SUBMISSION OF MANUSCRIPT Authors should submit manuscript electronically either as a PDF or Microsoft Word file to the Editor-in-Chief, or one of the Regional Editors, or one of the Associate Editors listed above. Authors are highly encouraged to submit manuscript electronically to save time for the reviewing process. ADVANCED SCIENCE LETTERS VOLUME 21, NUMBER 10 2947–3429 (2015) OCTOBER 2015 ISSN: 1936-6612, EISSN: 1936-7317 www.aspbs.com/science EDITORIAL BOARD Filippo Aureli, UK Marcel Ausloos, Belgium Martin Bojowald, USA Sougato Bose, UK Jacopo Buongiorno, USA Paul Cordopatis, Greece Maria Luisa Dalla Chiara, Italy Dionysios Demetriou Dionysiou, USA Simon Eidelman, Russia Norbert Frischauf, Austria Toshi Futamase, Japan Leonid Gavrilov, USA Vincent G. Harris, USA Mae-Wan Ho, UK Keith Hutchison, Australia David Jishiashvili, Georgia George Khushf, USA Sergei Kulik, Russia Harald Kunstmann, Germany Alexander Lebedev, Russia James Lindesay, USA Michael Lipkind, Israel Nigel Mason, UK Johnjoe McFadden, UK B. S. Murty, India Heiko Paeth, Germany Matteo Paris, Italia David Posoda, Spain Paddy H. Regan, UK Leonidas Resvanis, Greece Wolfgang Rhode, Germany Derek C. Richardson, USA Carlos Romero, Brazil Andrea Sella, UK P. Shankar, India Surya Singh, UK Leonidas Sotiropoulos, Greece Roger Strand, Norway Karl Svozil, Austria Kit Tan, Denmark Roland Triay, France Rami Vainio, Finland Victor Voronov, Russia Andrew Whitaker, Ireland Lijian Xu, China Alexander Yefremov, Russia Avraam Zelilidis, Greece Alexander V. Zolotaryuk, Ukraine HONORARY EDITORS Richard Ernst * , ETH Zürich, Switzerland Eric B. Karlsson ** , Uppsala University, Sweden Douglas Osheroff * , Stanford University, USA * Nobel Prize Laureate ** Member of the Nobel Committee for Physics 1987–1998 (chairman in 1998) EUROPEAN EDITOR Prof. Dr. Wolfram Schommers Forschungszentrum Karlsruhe, Institut für Wissenschaftliches Rechnen, D-76021 Karlsruhe, GERMANY Tel.: +49-7247-82-2432; Fax: +49-7247-82-4972; E-mail: [email protected]ASIAN EDITOR Dr. Katsuhiko Ariga, Ph.D. Advanced Materials Laboratory, National Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki 305-0044, JAPAN Tel.: +81-29-860-4832; Fax: +81-29-860-4832; E-mail: [email protected]Editor-in-Chief: Dr. Hari Singh Nalwa, Ph.D. American Scientific Publishers, 26650 The Old Road, Suite 208, Valencia, California 91381-0751, USA Phone: (661) 799-7200 Fax: (661) 799-7230 E-mail: [email protected]Diederik Aerts, Belgium Yakir Aharonov, Israel Peter C. Aichelburg, Austria Jim Al-Khalili, UK Simon Baron-Cohen, UK Jake Blanchard, USA Franz X. Bogner, Germany John Borneman, USA John Casti, Austria Masud Chaichian, Finland Sergey V. Chervon, Russia Kevin Davey, USA Tania Dey, Canada Frans de Waal, USA Roland Eils, Germany Marco Genovese, Italia Bert Gordijn, The Netherlands Thomas Görnitz, Germany Ji-Huan He, China Nongyue He, China Irving P. Herman, USA Dipankar Home, India Jucundus Jacobeit, Germany Yuriy A. Knirel, Russia Arthur Konnerth, Germany G. A. Kourouklis, Greece Peter Krammer, Germany Andrew F. Laine, USA Minbo Lan, China Martha Lux-Steiner, Germany Klaus Mainzer, Germany JoAnn E. Manson, USA Mark P. Mattson, USA Lucio Mayer, Switzerland Efstathios Meletis, USA Karl Menten, Germany Yoshiko Miura, Japan Fred M. Mueller, USA Garth Nicolson, USA Nina Papavasiliou, USA Panos Photinos, USA Constantin Politis, Greece Zhiyong Qian, China Reinhard Schlickeiser, Germany Surinder Singh, USA Suprakas Sinha Ray, South Africa Koen Steemers, UK Shinsuke Tanabe, Japan James R. Thompson, USA Uwe Ulbrich, Germany Ahmad Umar, Saudi Arabia ASSOCIATE EDITORS
165
Embed
ADVANCED SCIENCE LETTERS - ISTIEistie.org/21ASL10-pages-proofsneedcorrections.pdf · James R. Thompson, USA Uwe Ulbrich, Germany Ahmad Umar, Saudi Arabia ASSOCIATE EDITORS ... Pitak
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
SUBMISSION OF MANUSCRIPT
Authors should submit manuscript electronically either as a PDF or Microsoft Word file to the Editor-in-Chief, or one of the Regional Editors, or one of the Associate Editors listed above. Authors are highly encouraged to submit manuscript electronically to save time for the reviewing process.
ADVANCED SCIENCE LETTERSVOLUME 21, NUMBER 10 2947–3429 (2015)
OCTOBER 2015 ISSN: 1936-6612, EISSN: 1936-7317
www.aspbs.com/science
EDITORIAL BOARD
Filippo Aureli, UK
Marcel Ausloos, Belgium
Martin Bojowald, USA
Sougato Bose, UK
Jacopo Buongiorno, USA
Paul Cordopatis, Greece
Maria Luisa Dalla Chiara, Italy
Dionysios Demetriou Dionysiou, USA
Simon Eidelman, Russia
Norbert Frischauf, Austria
Toshi Futamase, Japan
Leonid Gavrilov, USA
Vincent G. Harris, USA
Mae-Wan Ho, UK
Keith Hutchison, Australia
David Jishiashvili, Georgia
George Khushf, USA
Sergei Kulik, Russia
Harald Kunstmann, Germany
Alexander Lebedev, Russia
James Lindesay, USA
Michael Lipkind, Israel
Nigel Mason, UK
Johnjoe McFadden, UK
B. S. Murty, India
Heiko Paeth, Germany
Matteo Paris, Italia
David Posoda, Spain
Paddy H. Regan, UK
Leonidas Resvanis, Greece
Wolfgang Rhode, Germany
Derek C. Richardson, USA
Carlos Romero, Brazil
Andrea Sella, UK
P. Shankar, India
Surya Singh, UK
Leonidas Sotiropoulos, Greece
Roger Strand, Norway
Karl Svozil, Austria
Kit Tan, Denmark
Roland Triay, France
Rami Vainio, Finland
Victor Voronov, Russia
Andrew Whitaker, Ireland
Lijian Xu, China
Alexander Yefremov, Russia
Avraam Zelilidis, Greece
Alexander V. Zolotaryuk, Ukraine
HONORARY EDITORS
Richard Ernst*, ETH Zürich, Switzerland
Eric B. Karlsson**, Uppsala University, Sweden
Douglas Osheroff*, Stanford University, USA*Nobel Prize Laureate **Member of the Nobel Committee for Physics 1987–1998 (chairman in 1998)
EUROPEAN EDITOR
Prof. Dr. Wolfram SchommersForschungszentrum Karlsruhe, Institut für Wissenschaftliches Rechnen, D-76021 Karlsruhe, GERMANYTel.: +49-7247-82-2432; Fax: +49-7247-82-4972; E-mail: [email protected]
ASIAN EDITOR
Dr. Katsuhiko Ariga, Ph.D.Advanced Materials Laboratory, National Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki 305-0044, JAPAN
ADVANCED SCIENCE LETTERSVOLUME 21, NUMBER 10 2947–3429 (2015)
OCTOBER 2015 ISSN: 1936-6612, EISSN: 1936-7317
www.aspbs.com/science
A SPECIAL ISSUE
2947–2951 Selected Peer-Reviewed Articles from the 3rd International Conference on Internet Services Technology and Information Engineering 2015 (ISTIE 2015)
Guest Editors: Ford Lumban Gaol and Benfano Soewito
Adv. Sci. Lett. 21, 2947–2951 (2015)
REVIEWS
2952–2956 A Review on Feature Selection Methods for Sentiment Analysis
Lai Po Hung, Rayner Alfred, and Mohd Hanafi Ahmad Hijazi
3385–3388 Ant Colony Algorithm and Its Application in the Fruit and Vegetable Wholesale Market in Vehicle Scheduling
Wei Xu, Liang Huang, and Hongren Wang
Adv. Sci. Lett. 21, 3385–3388 (2015)
3389–3391 Significance of Preparedness in Flipped Classroom
Azlina A. Rahman, Baharuddin Aris, Mohd Shafie Rosli, Hasnah Mohamed, Zaleha Abdullah, and Norasykin Mohd Zaid
Adv. Sci. Lett. 21, 3389–3391 (2015)
3392–3395 Building a Data Mart Using Single Dimensional Data Store Architecture with Student Subject: Case Study at Muhammadiyah University of Yogyakarta
Fajar Rianda, Asroni, and Ronald Adrian
Adv. Sci. Lett. 21, 3392–3395 (2015)
3396–3399 Contributing Factors of Online Brand Trust in Airline Industry
Nur Atika Jamuary, Mohd Shoki Md Ariff, Hayati Jamaludin, Khalid Ismail, Nawawi Ishak, and Mohd Sawal Abong
Adv. Sci. Lett. 21, 3396–3399 (2015)
3400–3404 Understanding Attitude of Online Shoppers: Integrating Technology and Trust Factors
Li Yuan Hui, Mohd Shoki Md Ariff, Norhayati Zakuan, Norzaidahwati Zaidin, Khalid Ismail, and Nawawi Ishak
Advanced Science Letters is a multidisciplinary peer-reviewed journal with a very wide-ranging coverage, consolidates fundamental and applied research activities by publishing proceedings from international scientifi c, technical and medical conferences in all areas of (1) Physical Sciences, (2) Engineer-ing, (3) Biological Sciences/Health Sciences, (4) Medicine, (5) Computer and Information Sciences, (6) Mathematical Sciences, (7) Agriculture Science and Engineering, (8) Geosciences, (9) Energy/Fuels/Environmental/Green Science and Engineering, and (10) Education, Social Sciences, and Public Policies. This journal does not publish general research articles by individual authors.
Research Topics Covered (but not limited to):
Advanced Science Letters deals with Adhesion Science and Technology, Aeronautics Engineering, Aerosol Science and Technology, Aerospace Engineering, Agriculture Engineering, Agriculture Sciences, Anthropol-ogy, Astronomical Sciences, Biochemical engineering, Biochemistry, Bioengineering, Bioinformatics, Biological Sciences, Biomedical Engi-neering, Biomedical Sciences, Biotechnology, Botany, Ceramic Science and Engineering, Cereal Chemistry, Chemical Biology, Chemical Engi-neering, Chemical Engineering, Chemistry, Civil Engineering, Clinical Sciences, Colloid Science, Communication Science, Composites Science, Computer Science, Engineering and Technology, Dairy Science, Device Engineering, Drug Discovery, Earthquake Science, Ecological Sciences, Educational Sciences, Electrical Engineering, Electronics Engineering, Energy Science and Technology, Environmental Engineering, Environ-mental Sciences, Enzyme Science and Engineering, Food Science, For-estry, Fuel Science, Genetics, Geosciences, Health Sciences, Hydrology, Information Technology, Interface Science, Life Sciences, Lubrication Science, Manufacturing Science, Engineering and Technology, Marine Science, Materials Science, Mathematical Sciences, Mechanical Engineer-ing, Medicinal Chemistry, Medicinal Science, Membrane Science, Metal-lurgical Science and Engineering, Meteorology, Microbiology, Minerals Science, Nanoscience, Nanotechnology, Nanoengineering, Nanomedicine, Nanobiology, Neuroscience, Nutrition Science, Oceanography, Optical Engineering, Optical Sciences, Paleontology, Paper Science, Petroleum Science, Petrology, Pharmaceutical Sciences, Pharmacology, Physics, Plant Sciences, Plasma Science and Technology, Polymer Engineering, Polymer Science, Polymer Technology, Powder Technology, Seismology, Sol-Gel Science, Supramolecular Science, Surface Science, Toxicology, Vacuum Science and Technology, Virology, Waste Management, Water Science, Wood Science and Technology, Zoology, Educational Aspects in all these Research Areas, and Selected Conference Special issues on Edu-cation, Social Sciences, Public Policies at the discretion of Editor-in-Chief.
Readership:
The journal is intended for a very broad audience working in all fields of (1) Physical Sciences, (2) Engineering, (3) Biological Sciences/Health Sciences, (4) Medicine, (5) Computer and Information Sciences, (6) Mathematical Sciences, (7) Agriculture Science and Engineering, (8) Geosciences, (9) Energy/Fuels/Environmental/Green Science and Engineering, and (10) Education, Social Sciences, and Public Policies. This journal does not publish general research articles by individual authors.
Manuscript Submission:
Submit Your Manuscript Electronically as a PDF or MS Word file to the Editor-in-Chief or European Editor or Asian Editor or one of the Associ-ate Editors.
Editorial Office:
ADVANCED SCIENCE LETTERSAmerican Scientific Publishers26650 The Old Road, Suite 208Valencia, California 91381-0751, USATel.: (661) 799-7200 Fax: (661) 799-7230E-mail: [email protected]
Books for Review:
Publications should be sent to the Editorial Office.
Biotechnology and BioEngineering Abstracts, Biotechnology Research Abstracts, Bacteriology Abstracts (Microbiology B), Neurosciences Abstracts, Engineering Research Database, Technology Research Data-base, Environmental Science and Pollution Management.
Referee’s Report:
Please prepare and submit Reviewer’s Report to appropriate Associate Editors accordingly.
Subscription & Advertisement:
American Scientific Publishers26650 The Old Road, Suite 208Valencia, California 91381-0751, USATel.: (661) 799-7200 Fax: (661) 799-7230E-mail: [email protected]
Annual Subscription Rates (Print Edition):
2015, Volume 21Institutional: US$ 3440 (USA), US$ 3440 (Foreign)Postage and handling charges: US$ 300 (USA), US$ 600 (Foreign).The publisher reserves the right to refuse nonqualified subscriptions.
Web Edition:
Advanced Science Letters will be available via internet. For subscription rates to Web Edition, please contact publisher.
Honorable Claim Period:
Claims for issues not received will be honored only if submitted within 60 days of the issue date for subscribers in North America or 120 days for all other subscribers. Send your written request to American Scientific Publishers.
Digital Object Identifier (DOI):
The DOI identification system for digital media has been designed to provide persistent and reliable identification of digital objects. Information on the DOI and its governing body, the International DOI Foundation, can be found at http://www.doi.org. The DOI appears in the lower right of the first page of each article.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form whatsoever by any means (electronic, mechanical, recording, photocopying, scanning), or translated into any foreign language or otherwise without the written permission of the Publisher. Only single copies of complete or partial articles may be made for personal and internal research use as allowed by national copyright laws; however, for copying beyond this the copier pays the stated per copy fee through the Copyright Clearance Center, Inc. (CCC). The registered trademarks, names and similar related mate-rials used in this journal are not to be considered unprotected by law.
Although this journal is carefully produced, the authors, editors, and pub-lisher do not guarantee the information and material contained herein to be free of errors. Contributed statements and opinions expressed in Advanced Science Letters (articles, communications, reviews, and research news) are those of the individual contributors and do not necessarily reflect the opinions of American Scientific Publishers, and its Editors assume no responsibility for them. American Scientific Publishers assumes no responsibility or liability whatsoever for any damage or injury, losses, or costs of any kind to person or property that arise due to the use of any materials, instructions, statements, opinions, methods, procedure, infor-mation, advertising materials, or ideas contained herein, negligence or oth-erwise. American Scientific Publishers expressly disclaims any implied warranties of merchantability or suitability for a particular purpose.
Copyright and Reprint Permissions:
Authorization to photocopy is granted by the ASP, provided that the appropriate fee is paid. Prior to photocopying, please contact the Copyright Clearance Center, Customer Service, 222 Rosewood Dr., Danvers, MA 01923, USA; +1 (508) 750-8400. For all other copying such as distribution, promotional, advertising, sale, collective work per-mission, write to: American Scientific Publishers, 26650 The Old Road, Suite 208, Valencia, California 91381-0751, USA.
Department of Electrical Engineering, Faculty for Engineering, King Mongkut’s Institute of Technology Ladkrabang,1 Chalongkrung Road, Ladkrabang, Bangkok, 10520, Thailand
An electromagnetically power transmission system to transmit power at a high power density and with a highefficiency under consideration of EMC aspects is proposed. Product requirements like size and weight aretaken into account. The system uses the resonance wave coupling and an impedance adaption for multi-stageconversion is implemented to optimize efficiency and impedance matching. The transferred power is up to 180 Wat a frequency of 6.78 MHz with a distance between 30 mm and 80 mm. The results show that an efficiencyup to 88% can be achieved. The transmitted power density reaches up to 14 W/cm2. It is shown that dueto the evanescent wave resonance coupling the system emits only very little radiation and thus is capable tokeep EMC/EMF regulations. Wireless power transfer applications such as electric vehicle charging and wirelesspower drive for LED lighting through a wall are demonstrated.
Keywords: Wireless Power Transmission, High Efficiency, Power Density, EMC, EMF, Ferrite Material.
1. INTRODUCTIONWireless power transfer is useful where connection cables are
inconvenient, dangerous or even impossible. But the major disad-
vantages of existing systems are1–3 low efficiency, typical <40%,
and low power density as 0.004 W/cm2, large and heavy equip-
ment and only low amount of power transmissible. In most
cases it is only possible to keep conformity according to EMC
and EMF regulations with huge efforts. This research proposes
energy transfer by evanescent wave resonance coupling with high
power density. Advantages are a smaller size of the system com-
ponents together with higher efficiency. Due to the methodology,
biological side effects can be neglected; the system complies
with the International Commission on Non-Ionizing Radiation
Protection (ICNIRP) regulations. The principle of function is the
evanescent wave coupling like described in Refs. [4, 5]: Basi-
cally using a set-up, consisting of two wire coils each with a
diameter of 60 cm, one coil as a transmitter and a second coil
which transforms the energy to the load (a bulb). The system
there is able to transmit power over a distance of two meters. The
coils resonating at 10 MHz use the resonant wave coupling to
avoid that the energy radiates uncontrolled through the air. In the
first demonstration, the researchers showed that the set up can
transfer power with an efficiency of 45%. The particular con-
cept in this research here is the combination of the relative low
operating frequency of 6.78 MHz, a selected ferrite material that
∗Author to whom correspondence should be addressed.
has its resistive contribution of the complex permeability above
the operation frequency and an optimized impedance matching
between all stages and interfaces with an advanced transmitter
inductor concept to achieve a high efficiency.
2. PROPOSED PRINCIPLE OF HIGH POWER
DENSITY WIRELESS TRANSMISSION
SYSTEM2.1. Resonance Coupling System
The resonance coupling effect is based on the evanescent wave
coupling which provides advantages like almost no stray field
to achieve a high efficiency, to fulfill EMF/EMC requirements
and to achieve a higher distance compared to other mecha-
nisms. In the previous work of André Kurs about wireless power
transfer,4�5 investigations have shown that efficiency in the power
transfer rapidly decreases with increasing distance. Therefore it
is necessary to use large coils with a diameter of 60 cm in
order to achieve longer transmission distance. Furthermore, the
impedance and thus the resonance frequency of the receiving coil
is very likely to be influenced by the load. Thus it is necessary
to minimize this effect being independent to supply any kind of
load. The system of this work composes of two main parts: First
part is the transmitter and the second part is the receiver, the
block diagram is shown in Figure 1.
The transmitter, with a resonance circuit operating at
6.78 MHz is supplied by a power signal generator. The receiver
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 2967–2973, 2015
DistancePowersignalgenerator
Trans-mittercircuit
Receivercircuit
Matchingtransfor-mer
Rectifier,Stabilizasion
Overvoltagelimiter
DC/DCconverterLoad
Fig. 1. Transmitter and receiver system set up and block diagram.
module consists of a resonance circuit and a diode bridge rectifier
to convert the RF voltage into a DC voltage. To buffer the energy
against current dips a bulk capacitor stabilization is added, fol-
lowed by a voltage limiter to protect the following circuitry from
over voltage. A DC/DC buck converter provides a regulated out-
put voltage in the range between 5 V and 24 V where any load
can be connected.
2.2. Set Up of the Transmitter and Receiver Circuit
2.2.1. Transmitter Circuit
The transmitter is set up as a series resonance circuit, to provide a
stable behavior of the power amplifier due to the low impedance
at resonance. The transmitter coil has 5 turns, which results in an
inductance of 1.58 �H at a frequency of 6.78 MHz with the cho-
sen ferrite material. The necessary series resonant capacitance,
shown in Figure 2, is in accordance to Eq. (1).
XL = XC
�L= 1/�C
C = 1
L�2= 1
1�58×10−6 · �2� ·6�78×106�2≈ 349pF
(1)
For achievement of a high efficiency it is essential to use a
proper ferrite material6�7 for the inductors and not an air coil.
A prototype of a wireless power transfer system is built in
Ref. [8]. The contribution of the energy in the transmission reso-
nance system must be primarily magnetic. This is achieved by a
high inductance and a low capacitance. The ferrite characteristics
give the advantage to reduce the circuit in size and to increase the
efficiency because of a field concentration and a proper adaption
to the field impedance. The ferrite material has its resistive part
of the complex permeability above 8 MHz and thus keeps the
losses low. The chart of the permeability is shown in Figure 3.
The material is selected in such a manner that at the operating
frequency the imaginary contribution (u′′s ) of the complex perme-
ability is almost zero what keeps the losses inside the ferrite at a
minimum.9
Transmission InduktorL = 1.58uH
Power Input
C = 349 pF – Cx/2 Cx: Variable capacitorfor fine-tuning
Fig. 2. Transmitter resonance circuit.
us`= 125
us``~ 1 at 6.78 MHz
Resistive(losses)
Reactive
Fig. 3. Ferrite characteristic: Complex permeability as a function offrequency.7
The complex permeability represents the “magnetic behavior”
of the ferrite material and has two contributions, the reactive
portion (u′s), which represents the inductance shown in Eq. (2),
and the resistive portion (u′′s ), which represents the losses shown
in Eq. (3). The initial permeability of the ferrite material is 125
as shown in the diagram.
LS = Lo ·u′s (2)
Rs = ��L� ·u′′s (3)
Then, the total impedance is
Z = �j�L� · �u′s −u′′
s � (4)
The absorption factor can be defined as
tan= Rs/�Ls = �u′′s �/�u
′s� (5)
With implementation of a ferrite core the losses in terms of
a resistive component add and the simplified diagram, without
parasitic effects is shown in Figure 4.
In case of an alternating magnetization of the ferrite the flux
density B is not in phase with the magnetic field produced. In
case of “small” magnetization, the angle between the result-
ing magnetization and the flux density represents the loss angle
shown in Eq. (5) and Figure 5.
The smaller the angle the lower are the losses and the higher
is “quality” of the material.9�10 The over-all loss angle of the
inductance is a combination of the contribution of the ferrite and
the wires of the coil. In case of larger magnetization the main
losses of the ferrite are the hysteresis, eddy current and magnetic
creep losses. The high specific resistance of material, which is
NiZn-ferrite shown in Table I, is required to reduce the eddy
currents losses.
Fig. 4. Ferrite core loss components, stray capacitances neglected.
2968
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 2967–2973, 2015
Fig. 5. Ferrite core loss parameters.
With rising frequency, current does not flow homogeneous
through the entire cross-section of the conductor but is more con-
centrated at the surface. The higher the frequency, the more cur-
rent is concentrated on the surface. This results in higher I2 ∗Rlosses and thus energy loss with rising frequency, the current den-
sity varies exponentially as a function of depth from the surface
of the wire. The effect is called skin effect. The skin depth is
defined as the distance below the surface where the current den-
sity has fallen to 1/e or 37% of its value at the surface.9 The skin
depth at 6.78 MHz for a copper wire is calculated and expressed
in Eq. (6).
=√2/��= 1√
�f�0�r(6)
�: Angular frequency of current (2�f in 1/s)
�: Magnetic permeability (�r ·�o� : Conductivity of the material (S/m or 1/�m).
= 1�05 1/�m
�0 = 4� ·10−7 H/m
cu = 5�82 ·107 �/m
=√
2
2� ·6�78 ·106 ·4� ·10−7 ·5�82 ·107 ·1�05= 24�7 �m
The parameter above show possible high losses due to the
skin effect. A similar effect, the Eddy current losses in the wind-
ings cause a main contribution of the losses. They are depending
on following parameters of Eq. (7), whereas the material and
the field must be uniform and the skin effect may here not be
considered:11
P = �2B2pd
2f 2
6k�D(7)
P : Power lost per unit mass (W/kg),
Bp: Peak of the magnetic field (T),
D: Diameter of the wire (m),
F : Frequency (Hz),
k: A constant equal to 1 for a thin sheet and 2 for a thin wire,
�: Resistivity of the material (� m), and
D: Density of the material (kg/m3).
This equation is valid only under the so-called quasi-static con-
ditions, where the frequency of magnetization does not result in
the skin effect what means the electromagnetic wave fully pen-
etrates the material which is the case at very low frequencies.
Table I. Specific resistance of selected materials.
Composition Specific resistance at T = 25 �C (�m)
MnZn-Ferrite 0.1–10NiZn-Ferrite 105–106
Ferrite
Strands, twisted
Fig. 6. Litz wire of the transmitter ferrite core.
But it shows clearly the dependency of the wire parameters like
thickness (d) and material constants (� and D) which certainly
do influence in case of higher frequencies, too. The other param-
eters are system depending, cannot be varied and thus not be
optimized.
Another factor, which may influence the efficiency, is the cur-
rent density, which is the ratio of current intensity to the area,
perpendicular to current direction, through which the current is
flowing. The mathematical definition of current density, which is
applicable to any possible distribution of charges flowing in the
conductor is11
I =∫s
�J ·d �S (8)
where �J is the current density at the area element d �S, and I is
the total current through area [A/m2]. In case one wire/strand
with a diameter of � 0.8 mm is used for the transmission coil,
a current density of up to 65�4∗106 A/m2 is caused, which is far
too much. The current density limitation for standard applications
is 1�0∗106 A/m2, which is about 65 times less! The construction
method of the transmitter coil is essential for the high efficiency
of the system. The assembling of a coil is depending on several
parameters: Number of turns, material of wire, diameter of wire
and for the litz wire the number of wires and the method of
twisting. The following picture (Fig. 6) shows the litz wire on
the core, which is eventually used.
Another important parameter of the transmission resonance
system is the self-resonant frequency of the coil which is a
parallel resonance consisting of the inductance and the parasitic
capacitances between the windings of the coil. As the series
capacitance of the resonance circuit is quite small, the impact
of the parallel capacitance must be considered and kept small.
C1
111 pF
10:3
L2
P SL3
Fig. 7. Receiver resonance circuit with decoupling transformer.
2969
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 2967–2973, 2015
It is necessary to use a suitable wire size and a proper layout.11
As mentioned in terms of the current density the variation of
the coil may change with the number of turns, twisting fac-
tor, amount of litz wires, density of turn layout, the impedance
matching to power amplifier and to the system receiver and the
magnetic field contribution of the efficiency.
2.2.2. Set Up of the Receiver Circuit
The receiver resonance circuit has to supply a nonlinear load
like the rectifier bridge, bulk capacitors and the DC/DC converter
and has to match to the near field wave impedance conditions.
Therefore, a parallel resonance circuit has been chosen. For opti-
mizing the decoupling between the resonance circuit and rectifier
bridge and to increase the efficiency by impedance matching a
decoupling transformer has been implemented (Fig. 7).
The parameter of the inductance and the capacitance can be
calculated based on the same resonant frequency, 6.78 MHz
Power meter Input / Output wave
Power meterheads
Poweramplifier
Signalgenerator
Test set-up measuring system
Test set-up diagram
Dirctionalcoupler
Voltage/Current meters
Transmitter,Receiver
Receiver
Transmitter
(a)
(b)
Fig. 8. Test set-up for the precise measurements of the system efficiency.
Reflected signal from load (Channel B): 130 mVss
Signal from Generator(Channel A):920 mVss
Fig. 9. Signals at directional coupler (50 ns/div, 200 mV/div).
but with consideration of the transformed additional impedances.
With the implementation of the additional decoupling, resp.
matching transformer the AC/DC power for any nonlinear load
can be supplied.
3. TRANSMISSION PERFORMANCE OF THE
SYSTEMThe centered diameter of the core is 42 mm, thus the efficiency
is measured with a gap of 42 mm as set-up in Figure 8. Figure 9
illustrates the signals decoupled by a 40 dB directional coupler
at the oscilloscope.
The upper trace is the reflected voltage from load at output B
(Fig. 8(a)), about 130 mV and the lower trace is the signal from
preamplifier at output B with 920 mV at an operating frequency
of 6.78 MHz (50 ns/div).
Result:
� After warm up (>1 h): core: 29 �C, coil: 32 �C,
Pin: 10.2 W, VDC: 22.5 V, RL: 56 Ohm
�% = �22�5�2
10�2×56·100 ≈ 88�6%
The transmitted power density is defined as the ratio between
the maximum transmitted power and the active area of the induc-
tor core (Fig. 10). At a transmitted power of 80 W with the
cross section of the ferrite core like shown in Figure 10 which
is 15�3× �102�4− 65�5� = 564�57 mm2 or 5.65 cm2 the maxi-
mum transmitted power density achieved is 0.14 W/mm2 which
is 140 kW/m2! For comparison, the power density of the sun
light is 1.37 kW/m2, the power density of Uran12 is 650 kW/m2.
4. SYSTEM LOSSES AND EMC/EMF
CONSIDERATIONSSystem losses are uncontrolled radiation and heat conversion.
The losses in the system compose of losses in the inductor (inner
and outer), losses in the capacitors and losses in the DC conver-
sion components. Some of those effects sum up and the energy is
converted into heat, others may radiate. The main contribution of
102.4 mm
65.5 mm
15.3 mm
Fig. 10. Cross section area of the transmitting and receiving core.
2970
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 2967–2973, 2015
Lowest level:0.1uT
–
Operating Frequency:6.78MHz
Fig. 11. ICNIRP limits of magnetic flux density exposure to human body.
the losses in the circuit are those of the inductors, the transmit-
ter capacitor is the second large contribution of the losses. The
losses of the inductors are contributions of the ferrite material
and of the wire windings. The primary losses of the ferrite are the
hysteresis-, eddy current- and magnetic creep-losses, the losses
of the wire windings are resistive losses and eddy current losses
due to the skin- and the proximity effects.11 The proper selec-
tion of the ferrite material and the optimal construction of the
wire winding is essential to reduce the losses and thus to reach
an optimum of efficiency.13 The losses in the set up are mainly
impedance matching losses between the power transmitter and
the transmission resonance circuit.
For keeping the EMC requirements the CISPR 1114 standard
for industrial, scientific and medical (ISM) equipment is consid-
ered. CISPR 11 defines ISM-designated frequency bands, which
are exempted from emission requirements. This means that there
are no radiation limits at certain frequency ranges. One range
defined in CISPR 11 is from 6.765 MHz to 6.795 MHz with a
center frequency of 6.780 MHz.
Today it is a must for a state of the art product to consider the
biological effect on human body. The International Commission
on Non-Ionizing Radiation Protection (ICNIRP) publishes guide-
lines for limiting RF exposure that provides protection against
bigcore
Strayfield
Sensors
Fig. 12. Set up of radiation measurement with a high sensitivity sensor.
Spectrumanalyzer
Field probes
Transmission systemRF source
Loopantenna
Powermeters
Load
Fig. 13. Set up of the wireless power transmission system to evaluate theEMC radiation.
known adverse health effects.15 In this work it can be shown
that the field used to transfer the power is limited to the area
between transmitter and receiver and thus outside of the beam
has less biological effect. The set-up can even keep the limit level
of 0.1 �T , which is the lowest level in the ICNIRP limit chart
starting at a frequency of 10 MHz and going up to 300 MHz as
shown in Figure 11.15
With a high sensitivity sensor, detecting fields in the range of
0.05 uT, it can be shown that the field of the inductance is limited
to the area between transmitter and receiver and only susceptible
to receivers at the same resonance frequency and thus outside of
the beam has no considerable biological effect. Figure 12 illus-
trates the set-up.
For further investigation, the system is set up like shown in
Figure 13. The RF source, an amplifier of 20 W is put in a
shielded box to avoid any unwanted radiation. The transmitter
and the receiver are placed on a non-metallic box which has no
influence on the electromagnetic behavior of the system. At a dis-
tance of 1.2 m a magnetic loop antenna is placed and connected
Power beam
Transmitter coil Receiver coil
Center
Fig. 14. Measurement points where H-field and E-field emission iscaptured.
2971
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 2967–2973, 2015
Table II. Result of the emission measurements according to the set
up shown in Figure 13.
Level in Level inLoad Kind of Level at Level at power center of Antennacondition field transm. Rec. beam core level
High (15 W) electric 102 90 102 102 67Low (5 W) electric 102 91 102 102 67High (15 W) magnetic 74 78 116 84 67Low (5 W) magnetic 82 80 94 85 67
Note: All results in dBuV (relative).
to a spectrum analyzer to measure the magnetic field emission of
the system in dependence of different operating conditions.
At close distance to the transmitter and the receiver the emis-
sion of the electric field and the magnetic field is measured
with field probes. The measurements are done in the vicinity of
the ferrite cores, in the middle of the core system and directly
between the poles (Fig. 14).
The emission is in fact, besides of the heat losses, another
contribution of system losses, which decreases the efficiency of
the system. The radiation losses can be separated into classical
losses, which are those, radiated from cables and components
because of mismatch and parasitic impedances and in losses,
which occur because of the power transmission beam and actu-
ally should be, due to the nature of the resonance evanescent
wave-coupling, zero. Table II lists the result of the measurements.
Two load conditions are measured, one with 5 W power con-
sumption another with 15 W. At both load conditions the electric
and the magnetic field are measured at the vicinity of the trans-
mitter and the receiver, in the power beam and at the center of
the transmission core. Additionally the radiation of the magnetic
field is measured with a magnetic loop antenna. Figure 15 shows
the signal received by the loop antenna in low power operating
condition.
The results listed in Table II allow following important
conclusions:
(1) The electric field is at the transmitter very high due to the
high operating voltage at the capacitor. It surrounds the whole
transmission system with the same high level.
(2) The magnetic field is highest in the power beam and
decreases rapidly in the center of the cores. There is also a low
level surrounding the cores caused by stray fields of the coils.
Fig. 15. Emitted signal measured with the loop antenna, operating condi-tion: 5 W load at the receiver.
Fig. 16. Transmitting the power through the cement wall, 9 cm thickness.
(3) The magnetic field measured with the Loop antenna is inde-
pendent from the load (!). It shows that the power transmission
is restricted to the area between the cores.
The remaining emission is mainly caused by
—Radiation from the cable between the power amplifier and the
transmitter, due to mismatch and endless shielding effectiveness.
—Parasitic impedances of the transmitter and receiver resonance
circuits.
—Proximity and skin effect of the coils.
In conclusion when the receiver is taken away from the sys-
tem the remaining emission level at the loop antenna decreases
only to approx. 65 dBuV. Due to the high quality factor Q of
the transmitter resonance circuit the emission of any harmonics
caused by the RF power amplifier is very low what gives the
possibility to implement a hybrid class D RF-amplifier in small
size as a next application.
5. APPLICATIONSThe application in Figure 16 shows a wireless power transmission
through a 9 cm cement wall.
A further application may also be for example a storage energy
charging system for cars or in medical appliances. Applications
of other researchers focus on areas such as: Wireless power
transmission with multi receivers in power supply system’s,16
wireless power and data link,17 wireless charging systems18 or
wireless power standardization for supply and charging of small
appliance.19 Each of the applications has its individual advantage
but none of them can hardly combine the key features which
are small size, proper distance between transmitter and receiver,
EMC consideration and high efficiency. The applications operate
in the range from 100 kHz up to 27 MHz so many not in the
ISM band, some reach an efficiency up to 90% but work at low
distance and on magnetic coupling basis only.
6. CONCLUSIONIn this work it is shown that with pre-defined parameters and
restrictions for practical use as a precondition, it is possible to
achieve a result, which can be used already for industrial design.
The work shows that it is possible to transmit high power via
the air keeping still a high efficiency, a low weight, a small size
and also EMC/EMF restrictions. The circuit enables a power
2972
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 2967–2973, 2015
transmission under conditions, which make the receiver safely in
use as no cables, or batteries need to be applied. Thus there is a
wide range of application possible like in medical appliance or
appliance for chemical industry where special precautions must
be taken.
The parameters of wireless power performance are the distance
between transmitter and receiver, the transmitting frequency and
the system impedances. The key for the principle is the resonance
coupling effect together with a high performance ferrite inductor
which allow to fulfill the preconditions like EMC/EMF and effi-
ciency. The various load conditions normally affect the matching
impedances and thus they have to be decoupled from the transmis-
sion system here realized with implementation of an additional
matching transformer with again a proper high efficiency ferrite
material. The efficiency is significantly depending on the trans-
mitter resonance circuit where the highest power of the whole
system is handled. The components used here have to be opti-
mized regarding low loss and small size. The maximum efficiency
of 88.6% is achieved at a gap of 42 mm with 9 W output power.
The maximum transmitted power density achieved is 0.14 W/mm2
which is 140 kW/m2! The component and material selection and
design are essential to reach a high performance transmission. The
further progress in this work is focused on how to maintain the
high efficiency wireless power transmission at various distances
and how to keep the high efficiency at longer distance.
References and Notes1. J. J. Dubray, Standards for a service oriented architecture (2003), http://
www.ebxmlforum.org/articles/eb for_20031109.html.2. D. H. Akehurst, Transformations based on relations (2004), http://heim.
ifi.uio.no/_janoa/wmdd2004/papers/akehurst.pdf.
3. A. Kurs, A. Karalis, R. Moffatt, J. D. Joannopoulos, P. Fisher, and M. Soljacic,Science 317, 83 (2007).
4. A. Karalis, J. D. Joannopoulos, and M. Soljac, Elesevier, Annals of Physics323, 34 (2008).
5. A. Karalis, J. D. Joannopoulos, and M. Sol-jacic, Elsevier Annals of Physics323, 34 (2008).
6. H. Zenkner, A. Gerfer and B. Rall, Trilogy on Inductors, 3rd edn., Swiridof-fVerlag (2009), pp. 140–143, ISBN: 3-934350-73-9.
7. Ferroxcube: Soft Ferrites and Accessories, http://www.ferroxcube.com, DataHandbook (2008), Vol. 29, pp. 7–13.
8. B. L. Cannon, J. F. Hoburg, D. D. Stancil, and S. C. Goldstein, IEEE Transac-tions on Power Electronics 24, 1819 (2009).
9. Meinke, Gundlach, Taschenbuch der HF-Technik Bände I–III, Hrsg. Von K.Lange, and K.-H. Löcherer, Springer-Verlag (1992), pp. B3–B4, E13–E14,H1–H4, 17–19, ISBN: 3-540-54714-2.
10. K. G. Kaschke and H. Gmb, Co. Rudolf-Winkel-Str. 6, 37079 Göttin-gen, Nickel-Zink-Kobalt-Ferrite, http://www.kaschke.de/fileadmin/user_upload/documents/datenblaetter/Materialien/NiZn-Ferrit/K251.pdf, datasheet.
11. H. Kaden, Wirbelströme und Schirmung in der Nachrichtentechnik, edited byH. W. Meissner, Springer-Verlag (1959), Vol. 111, pp. 59–88.
12. Energy density at WWW.Wikipedia.org/wiki/Sonnenenergie.13. H. Zenkner, A. Gerfer, and B. Rall, Trilogy on Inductors, 3rd edn., Swiridoff-
Verlag, pp. 60–81, ISBN: 3-934350-73-9.14. CISPR 11:—Industrial, Scientific and Medical (ISM) Radio-Frequency
Equipment—EM Dis. Char.—Limits and Methods of Measurement. CISPR11,IEC (2011).
15. ICNIRP Guidelines: Guidelines for limiting exposure to time-varying elec-tric, magnetic and electromagnetic fields (up to 300 GHz), http://www.icnirp.de/documents/emfgdl.pdf, p. 512.
16. TinekeThio, A Bright Future for Subwavelength Light Sources, American Sci-entist (2006), Vol. 94, pp. 40–47.
17. R. Puers, K. V. Schuylenberght, M. Catrysse, and B. Hermans, Wirelessinductive transfer of power and data, Analog Circuit Design, Springer (2006),pp. 395–414.
18. R. Hui, Comparison of Power Savings Based on the Use of Wireless ChargingSystems and Conventional Wired Power Adapters, City University of HongKong (2009).
19. Wireless Power Consortium, Transfer efficiency, http://www.wirelesspowerconsortium.com/technology.
Received: 9 October 2014. Accepted: 19 November 2014.
Dini Handayani1�∗, Hamwira Yaacob1, Abdul Wahab Abdul Rahman1,Wahju Sediono2, and Asadullah Shah1
1Department of Computer Science, Kuliyyah of Information Communication and Technology2Department of Mechatronic, Kuliyyah of Engineering, International Islamic University Malaysia Kuala Lumpur, Malaysia
In the recent years, more studies that aim to make computers understand, experience and respond to variousemotional states accordingly through computational models have been widely researched. Conversely, little hasbeen done to recognize the medium term of emotion, such as mood. Thus, in this study, a mood modeling wasproposed to recognize mood from a sequence of several emotional states. The input for the proposed moodmodel was derived in the form of electroencephalogram (EEG) signals, which were captured from five subjectsduring eyes close, and eyes open. Our analysis indicates that mood can be recognized either from eyes closeor from eyes open.
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 2997–3001, 2015
Table I. Taxonomical table on mood formulation.
No Author Measurement tools and data Emotion recognition Mood formulation
1 Katsimerou et al.13 Self assessment, audio-visual TheHUMAINE affective database
Dimension: Valence and arousal Moving average
2 Thrasher et al.10 Body gesture during listening music Dimension: Valence, energetic arousal,and calmness.
Logistic regression from bodygesture
3 Livne et al.11 Data of 3D pose tracking and motioncapturing
Discrete: Sad and happy Logistic regression videoanalysis
4 Metallinou and Narayanan14 Self assessment, audio and video fromUSC CreativeIT database
Dimension: Activation, valence anddominance
Average of video analysis
5 Hashemian, et al.12 Face reader, interaction with computer Discrete: Sad, happy and neutral Bayes classification
Based on our literature review as shown in Table I, some
research works were implement to recognize mood from video.
Three studies reported on how mood of a real subject is directly
recognized in a real-time.10–12 None of the study utilize electroen-
cephalogram (EEG) for mood recognizer. Hence, this motivates
us to explore more on this matter.
In this study, a computational model of mood were constructed
from a sequence of several emotional states, its temporal prop-
erties were analyzed and proposed it as a model to automatic
recognize the mood from EEG.
EEG is an imaging tool that captures electrical activation
occurrences which is used to monitor the brain condition.16
Brainwaves which are represented by EEG signals are commonly
ranged into four bands; delta (0.5 Hz–4 Hz) Characteristic of
deep sleep phases,17 theta (4 Hz–8 Hz) Drowsiness and fatigue
due to monotonous task,18 control of working memory process,19
alpha (8 Hz–13 Hz) Cognitive control,20 creative thinking21 and
beta (13 Hz–30 Hz) Alertness,22 phonological tasks.23 Each of
the frequency bands was observed as products of different brain
tasks.
3. METHODOLOGYThe methodology of this research will be done with the following
step:
3.1. Signal Acquisition
Four EEG electrodes (C3, C4, T3, and T4) were pasted on their
scalp, with the specific regions using the International 10–20 sys-
tem. The electrodes will be connected into the EEG head box to
enhance the signals.
3.2. Signal Preprocessing
Here, all the signals were filtered to exclude noises and unre-
lated artifacts. Finally, only one-minute activities in the recorded
signals were used for the training.
Fig. 1. Flowchart of the methodology.
3.3. Feature Extraction
Kernel Density Estimation (KDE) was applied as a features
extraction technique. 100 features were extracted from the signals
for each instance.
3.4. Classification
Multi Layer Perceptron (MLP) was selected for classifiers. The
training was done based on dimensional approach. The classes
were labeled based on the generalized values of valence and
arousal. The values were assigned based on quadrants in which
each emotion is located in the affective space model, as shown
in Table II.
Bialoskorski et al.,24 defined color for emotional state. Happy
emotional state is indicated as having positive level of valence
and high arousal with an orange color. Quadrant of positive
valence and low arousal represents the calm emotional state with
green color. Sad emotion is located at the quadrant of negative
valence and low arousal with blue color. Finally, fear emotion
corresponds to negative valence and high arousal with red color.
During the experiment, the data obtained from the subjects
were fear, sad, and happy emotion.
For the dimensional approach, the performances are consis-
tent. For valence dimension, subject three gets the highest with
94.93% accuracy. Subject two and subject four get the lowest
with 90.92% accuracy. With the mean (M) 92.04% and standard
deviation (SD) 1.67.
For arousal dimension, subject three gets the highest with
93.36% accuracy. Subject two get the lowest with 90.92% accu-
racy. With M 92.18% and SD 1.31. As shown in Figure 3.
Figure 4 shows the scatter plots of the basic emotional states.
It is distributed well within the expected quadrants as described
in ESM.
3.5. Mood Formulation
Subjects are instructed to do the resting states (eyes open and
eyes close), one minute for each condition. Two episode of an
emotional state of a user were collected with a total time for
each episode is one minute (n). Sample rate for each emo-
tions are 0.004 seconds (t). One emotion can be estimated from
Table II. Valence and arousal labels for emotions classification.
Labels
Emotions Valence Arousal
Happy 1 1Calm 1 −1Fear −1 1Sad −1 −1
2998
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 2997–3001, 2015
Fig. 2. The emotional states model (ESM).
the sample rate. However, the overall episode will refer to the
moods.
In this research both emotions and mood refer to emotional
state of the EEG signal from the subject. From the classifi-
cation phase, basic emotion and label of the emotions were
defined.
e= �v�a� (1)
For the every episode, the emotion vector e, corresponding to
the recognize time as follow:
et = �vt�at�� t = 1�2�3� � � � � n (2)
Assuming E as sequence emotions
E = �e1� e2� e3� � � � � en� (3)
Finally, mood formulation is measured based on the corre-
sponding mood function, as a proposed work from the previous
study.25
m= F �E� (4)
The proposed mood model is based on MA of emotions over
time. The mood formulation, were divide into valence and arousal
mood.
M = �Vm�Am� (5)
For mood of valence, the formulation as follows:
Vm = MA of Valence (6)
Vm = Vt+1 (7)
Vt+1 = Vt−1 +��Vt −Vt−1� (8)
Table III. Parameters for MLP.
Parameters Values
Number of hidden layer 1Number of nodes in hidden layer 30Mean-square error goal 0.1Activation function at hidden layer Tan-sigmoidActivation function at output layer Pure linier
Fig. 3. Accuracy of Subject’s Identification.
Where: Vt+1 = Valence for the next period, �= Smoothing con-
stant, Vt =Observed value of valence in period t, Vt−1 = Previous
valence.
Likewise, Eqs. (6)–(8) can be re-written for arousal as:
Am = MA of Arousal (9)
Am = At+1 (10)
At+1 = At−1 +��At −At−1� (11)
Where: At+1 = Arousal for the next period, �= Smoothing con-
stant, At =Observed value of arousal in period t, At−1 = Previous
arousal.
3.6. Mood Recognition
Mood recognition is derived from the most prominent emotion.
Thus, the individual emotion was map into mood directly and
takes the quadrant of the mood space that containing the major-
ity of emotion. This quadrant may then be a predictor of the
recognized mood.
Fig. 4. Emotions distribution for subject 3.
2999
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 2997–3001, 2015
Fig. 5. Subject 2 eyes open emotions.
Fig. 6. Subject 2 eyes open, mood of the emotions.
4. EXPERIMENTAL RESULTThe main interest of this research will be the correlation between
emotion and mood of the subject. In addition, it is important to
get a higher accuracy, to prove that using dimensional approach
(valence and arousal) are better than categorical approach, as
discussed in Ref. [16].
Thus, memory test for all subjects were constructed to see the
level of accuracy, either it can be accepted or rejected as shown
in Figure 3. Eyes close and eyes open signal will be compared
to get mood recognition result.
The eyes open emotional states for subject two are shown in
Figure 5. The subject has the dynamic emotion. Initially, the sub-
ject was in the sad emotion, change to fear and happy emotions.
Calm emotions also appear during the eyes close.
The get the mood of emotion, MA were calculated for both
valence and arousal. With the smoothing average is 1/60 seconds.
As depicted on the Figure 5, the shortest emotions is one second
from the total episode 60 seconds.
�= 1
60sec
As shown in Figure 6, the emotions of subject two were
changed. The prominent emotions are sad emotions. Here, the
mood states can be derived from the prominent emotion. How-
ever, calm and fear emotion were still appear. To get the most
prominent emotion, MA was recalculated from the mood value.
From the Figure 7 above, the most prominent emotion is sad.
Therefore the mood of subject two in eyes open is sad mood.
Fig. 7. Subject 2 eyes open, mood of the mood.
Table IV. Experimental results.
No Subject Action Emotion Mood 1 Mood 2
1 Subject 2 EC – Sad Sad
2 Subject 3 EC – Fear FearEO – Sad Sad
3 Subject 4 EC – Happy HappyEO – Happy Happy
4 Subject 5 EC – Calm HappyEO – Calm Calm
5 Subject 6 EC – Happy HappyEO – Calm Calm
EO – Sad Sad
The mood from eyes open will be compared to eyes close. With
the assumption, the mood recognition will be the same.
As a summary, we have Table IV as an experimental result.
It is consist of five subjects with eyes close and eyes open. The
moods of eyes close and eyes open are similar for subject two
and four. However, for subject three, five and six, the mood is
different. Even though the emotions are different, all the emo-
tions are in the same valence area. Therefore, subject three in the
negative mood, subject five and six are in the positive moods.
5. CONCLUSIONSIn this paper, computational models of mood that infer the long-
term emotional state of a person have been proposed. The emo-
tional state is generated from EEG brainwave. Two emotional
episodes are taken into consideration; (1) eyes close and (2) eyes
open. It is showed that a model could get the prominent emotions
from second order of moving average as a mood recognition for-
mula. For the future work, mood recognition will be under the
executive function task, and correlate the result with the resting
states.
In addition, it is expected that the refined models are being
able to properly capture the process that regulate the relationship
between recognized emotions and mood. Finally, valence and
arousal values are predicted quite satisfactorily from the proposed
models, yet their combination into mood is less precise.
Acknowledgments: We would like to express our sincere
gratitude to Norzaliza M. Nor for providing us with the data
and valuable feedback. This work is supported by Fundamen-
tal Research Grant Scheme (FRGS) funded by the Ministry of
Higher Education (Grant code: FRGS14-137-0378).
References and Notes1. R. W. Picard, Int. J. Hum. Comput. Stud. 59, 55 (2003).2. D. Papachristos, K. Alafodimos, and N. Nikitakos, Emotion Evaluation of Sim-
ulation Systems in Educational Practice (2012), pp. 1–7.3. C. Pimentel, Affect. Comput. Intell. Interact. 72 (2011).4. S. Alghowinem, R. Goecke, M. Wagner, G. Parkerx, and M. Breakspear,
Head pose and movement analysis as an indicator of depression, 2013Human Association Conference Affective Computing and Intelligent Interac-tion, Sepetember (2013), pp. 283–288.
5. B. Bostan, Entertain. Comput. 262 (2010), no. Idmi.6. I. Siegert, R. Böck, and A. Wendemuth, Cogn. Behav. Syst. 273 (2012).7. J. H. Janssen, E. L. V. D. Broek, and J. H. D. M. Westerink, User Model.
User-adapt. Interact. 22, 255 (2011).8. D. Hume, Organ. Behav. 258 (2012).9. F. D. L. T. Torre and J. F. Cohn, Handb. Face Recognit. 1 (2011).
3000
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 2997–3001, 2015
10. M. Thrasher, M. D. V. D. Zwaag, N. Bianchi-Berthouze, and J. H. D. M.Westerink, Mood Recognition Based on Upper Body Posture and MovementFeatures, Springer-Verlag, Berlin, Heidelb. (2011), pp. 377–386.
11. M. Livne, L. Sigal, N. F. Troje, and D. J. Fleet, Comput. Vis. Image Underst.116, 648 (2012).
12. M. Hashemian, A. Nikoukaran, H. Moradi, M. S. Mirian, and M. Tehrani-Doost,Determining mood using emotional features, 7’th Int. Symp. Telecommun.,September (2014), pp. 418–423.
13. C. Katsimerou, J. Redi, and I. Heynderickx, A computational model for moodrecognition, 22nd Int. Conf. UMAP 2014, Aalborg, Denmark (2014), Vol. 8538,pp. 122–133.
14. A. Metallinou and S. Narayanan, Annotation and processing of continuousemotional attributes: Challenges and opportunities, Proc. 2nd Int. Work. Emot.Represent. Anal. Synth. Contin. Tome Sp. (Emosp. 2013), Shanghai, China(2013).
15. H. Yaacob, I. Karim, A. Wahab, and N. Kamaruddin, Two dimensional affectivestate distribution of the brain under emotion stimuli (2012), pp. 6052–6055.
16. H. Yaacob, Classification of EEG signals using MLP based on categoricaland dimensional perceptions of emotions, 2013 5th Int. Conf. Inf. Commun.Technol. Muslim World, March (2013), pp. 1–6.
17. K. Šušmáková, Slovak Academy of Sciences 4, 4 (2004).18. B. T. Jap, S. Lal, P. Fischer, and E. Bekiaris, Expert Syst. Appl. 36, 2352
(2009).19. P. Sauseng, B. Griesmayr, R. Freunberger, and W. Klimesch, Neurosci. Biobe-
hav. Rev. 34, 1015 (2010).20. B. Zoefel, R. J. Huster, and C. S. Herrmann, Neuroimage 54, 1427
(2011).21. A. Fink, B. Graif, and A. C. Neubauer, Neuroimage 46, 854 (2009).22. J. Kaminski, A. Brzezicka, M. Gola, and A. Wróbel, Int. J. Psychophysiol. 85,
125 (2012).23. B. Penolazzi, C. Spironelli, C. Vio, and A. Angrilli, Behav. Brain Res. 209, 179
(2010).24. L. S. S. Bialoskorski, J. H. D. Westerink, and E. L. V. D. Broek, Mood
swings: An affective interactive art system, ICST Institute of ComputerScience Social Informatics Telecommunication Engineering 2009 (2009),pp. 181–186.
25. D. Handayani, H. Yaacob, A. W. Abdul Rahman, W. Sediono, and A. Shah,Systematic review of computational modeling of mood and emotion, 5thInternational Conference Influence Communication Technology Muslim World,November (2014), pp. 1–5.
Received: 15 October 2014. Accepted: 29 November 2014.
Department of Informatics, Faculty of Industrial Technology, Universitas Islam Indonesia,Yogyakarta 55584, Indonesia
This paper is mainly discuss about Agile software development and strategy to build a successful agile team fordeveloping high quality software products. This paper starts with general information about agile developmentand two most popular methods, Extreme Programming (XP) and Scrum. XP has five core values and twelvebest practices inspired by the Agile Manifesto with the main goal is to organize software engineers so that theyable to produce software products with higher quality within short functionality delivery schedules. Scrum isan agile method which use a software development framework called sprint with a small, self-organizing andempowered team with rapid adaption and complete visibility. Then we will move onto the detail informationabout characteristics of Agile projects which are blurring roles, continuous development activities, and teamaccountability. The next part of this paper discuss about four key factors that have to be concerned by softwareengineers for making a successful Agile team. The key factors are co-located team, engaged customers, self-organizing team, and accountable and empowered team.
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3002–3006, 2015
Fig. 1. Agile manifesto.4
software product early and often to get feedback from customers
and the team itself to improve the quality of the software product.
• Respect: In a XP team, every software engineer contributes
value. He/she gives and feels the respect as a valued team mem-
ber. Software engineers and the customers respect each other in
term of their expertise. Also, management respects the responsi-
bility and the authority of the software engineers over their own
work.
• Courage: During the project, software engineers have to tell
the truth about their progress on tasks assigned to them. There
is no excuses for failure in a XP team since it plan to succeed.
But, software engineers in a XP team should not fear anything
because they work together and they will adaptively respond to
changes. We will adapt to changes when ever they happen during
the project.
XP has 12 practices which can be used by software engineers
to enact the above values in a software development process:7
• On-site customer
• User stories
• Metaphor
• Simple design
• Coding standard
• Pair programming
• Collective ownership
• Testing
• Refactoring
• Small releases
• Continous integration
• 40-hours workweek
Figure 2 is a flowchart which describes about how XP’s rules
work together during a software development process.
Scrum is an agile method which use a software development
framework called sprint. A Scrum project usually has a small,
Fig. 2. How extreme programming’s rules work together.8
self-organizing and empowered team with rapid adaption and
complete visibility. There are three main roles in Scrum which
are product owner, Scrum master, and developmet team. Product
owner in Scrum is like on-site customer in XP, he/she is rep-
resenting the interests of customers. His/her main responsibility
is to define and prioritize the requirements of the system being
developed. The scrum master is the “manager” of the develop-
ment team and responsible to ensure that the development team
effectively perform to achieve the goals of the project. The scrum
master also responsible for teaching scrum to software engineers
on the team so they can enact scrum values and practices. He/she
also must be able to resolve any problems during the project.
The development team consist of professional software engineers
who are responsible for implementing the system. The develop-
ment team in a Scrum project has uniques characteristics which
are cross-functional and self-organizing. The development team
does the actual work of delivering the product increment. All
software engineers in the development team should be available
to the project full time.5
In a Scrum project, there is a series of sprint which is a
2–4 weeks iteration to produce the tangible and tested artifacts.
Following are activities for each sprint in Scrum:9
• Product backlog review: this activity is performed by the
development team together with product owner and scrum mas-
ter by reviewing the product backlog. Product backlog is a list of
user stories representing the required functionalities of the system
being developed.
• Product backlog refinement: this activity is used by the devel-
opment team to discuss with product owner in order to change the
order of the user stories listed in product backlog, remove unnec-
essary requirements, add new functionalities, split and merge user
stories, and determine the user stories that will be finished on the
intended sprint.
• Sprint planning: it is a meeting to produce a list of user stories
will be completed in the sprint and a plan for finishing all related
tasks to complete the intended user stories.
• Sprint: during the sprint, software engineers produce the prod-
uct increment as be determined in sprint planning. Product incre-
ment is the most important deliverable in Scrum project. It has
to be high quality, meet the criterias of "done", and accept by the
product owner.
• Daily Scrum: in this daily the meeting, each of software engi-
neers have to explain about what he/she has finished, what he/she
plans for the next day, and what problems he/she has during
his/her work.
• Sprint review: this meeting is held at the end of each sprint.
On this meeting, the development team demonstrates the product
increment they finished on that sprint to the product owner. Then
the product owner will evaluate whether the product increment
is acceptable or not.
• Sprint retrospective: this activity is performed by the develop-
ment team to review their performance during the last sprint. The
development team discuss to identify potential improvements.
The framework which illustrated how Scrum project being per-
formed is shown in Figure 3.
2. CHARACTERISTICS OF AGILE PROJECTSThere are some special characteristics of Agile projects that
software engineers need to know about it. Understanding the
3003
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3002–3006, 2015
Fig. 3. The scrum framework.10
differences between Agile projects and projects that use other
“traditional” software development methods is a compulsory step
for software engineers before they start their Agile projects.11
The first and the most important characteristic of Agile
projects that distinguish them with non-Agile projects is that
roles are really blur on them.12 Titles or roles (Developer, Tester,
System Analyst, Database Engineer, User Experience Expert, etc)
are not important on Agile projects. People in an Agile project
must pitch in and do whatever it takes for the successfulness of
the project, regardless of their roles. This situation is illustrated
on Figure 4.
This characteristic doesn’t mean that an Agile project employ
people without specific requirements. Of course people joining
an Agile team must have core competencies and also need to
stick to what they are good at. But on an Agile project there are
no “narrowly defined” roles such as programmer, analyst, tester,
and so on and so forth.
The second characteristic of Agile projects is that they perform
all software development stages (i.e., analysis, design, coding,
and testing) as continues activities.13 In every single iteration,
an Agile team performs analysis, design, coding, and testing
activities for a little number on requirements. In the next iter-
ation it performs those activities again for other requirements.
Those activities are never end until the team successfully deliver
the intended system to the users. This situation is illustrated in
Figure 5.
Fig. 4. Roles on agile projects.12
Fig. 5. The difference between activities on traditional projects and agileprojects.12
From Figure 5 shown above we can see that software devel-
opment activities on Agile projects can’t exist in isolation any-
more. So, all software engineers in an Agile team must be able
to working together daily and perform those continues activities
throughout the project.
The third characteristic of Agile project we need to consider
is that quality is a “one team” responsibility.14 Every people
working on an Agile project is a “Quality Assurance Supervi-
sor” an the team, whether they managing the project, eliciting the
requirements, designing the user interface, or writing and testing
the code. Everyone must contribute for the team accountability.
The different situation is happen in traditional projects where the
quality of the requirements is system analyst’s responsibility, the
quality of the code is the programmer’s responsibility, the quality
of the user interface is the user experience’s responsibility, and
so on and so forth. This difference is illustrated on Figure 6.
3. KEY FACTORS FOR SUCCESSFUL
AGILE TEAM3.1. Co-Location
In term of the location of the people envolved in the project, soft-
ware development team can be divided into two types, co-located
team and distributed team. Both of them can run Agile projects.
But, have a co-located team is a better thing on an Agile project.
It can dramatically improve the productivity of the team.13
Effective communication plays an important role in Agile
projects and there is no better strategy to build a communica-
tive team rather than having everyone sit together in a co-located
workplace. On a co-located team, people can get answers for
their questions quickly, problems can be fixed on the spot, and
Fig. 6. The difference between accountability concept in Agile and tradi-tional methods.12
3004
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3002–3006, 2015
friction can be minimized. Also, intensive interactions among
team members allows trust to be built more quickly.14
Sometimes there is no choice but having a distributed team
to run the project. Following are two useful steps the distributed
teams can do to improve their productivity:13
• At the beginning of the project, team can reserved some bud-
get to bring everyone together even just for 1 or 2 weeks. Team
members must effectively use that time to know each other and
build “chemistry” among team.
• During the project, team can use possible communication
tools (Skype, video conference, social media) so it seem like a
co-located one.
3.2. Engaged Customers
Engaged customers are some stake holders of the intended sys-
tem who show up to system demos, answer questions about the
requirements, give suggestions and feedbacks, and provide the
guidance for the developer team to build a “right and working”
software.13 Basically they can be considered as members of the
developer team.
Having engaged customers that actively participate during the
project is very important. It is impossible for the team to build
compelling and innovative software without having people who
will use that software as part of the process.
All Agile methods fight hard for customers engagement
through practices. For example, Extreme Programming has it’s
on-site customer and Scrum has it’s dedicated role of product
owner.3
3.3. Self-Organizing
To be able to delivering a high quality software, an Agile team
need to performs as a self-organize team. It means that once the
team understands the goal of the project, every software engi-
neers on the team collectively figure out how to achieve that
goal.11
Self-organization is about how software engineers working
together as a team with their unique talents, skills, and passions,
no matter what their roles, so the team can best deliver the project
to its customers.
Self-organization can be considered as an acknowledgement
that the best way to make a successful team is to let the role
fit the person, not making the person fit the role.15 On an Agile
team, it is not a problem when a programmer involved in the
user interface design process or when a software tester involved
in the requirement elicitation process.
Following are some useful tips to get the team to self-
organize:14
• Team recruit people who capable to initiate ideas, have tech-
nical excellence and creativity, and don’t wait for instructions.
• Team must let everyone proposes ideas, creates the plan,
comes up with the estimates, and take ownership of the project.
• Team must worry less about roles and focus on the continuous
production of working and tested software.
Team must trust all members, encourage them, and empower
them to get the project done successfully.
3.4. Accountable and Empowered
A professional Agile team realizes that the customers are really
concern on the quality of the delivered software. So, everyone
Fig. 7. Daily stand up meeting.16
on the team must support the accountability of the results that
the team produces for the customers. Everyone on the team must
understand that the team has responsibility to deliver value for
the customers during the project.
To be able to accountable, the team must be empowered.
Empowering the team can be done by allowing all members
of the team make their own decisions, take their initiatives, do
what their think is right, and act on their own accord.12 It wil
encourage everyone to solve their own problems without wait for
permission from anyone.
One powerful strategy to maintain the accountability of the
team and to empower everyone on the team is get the team to
demo the software. Putting everyone in front of real customers
and having them to demo their works will effectively making the
team more accountable.15 First, everyone on the team will realize
that the real customers (with real problems) are counting on them
to deliver a right and working software. Second, getting the team
to demo the software to the real customers will be very useful to
collect the feedback that needed to improve the software. Daily
stand up meeting as shown in Figure 7 is an example of activities
that support the accountability of an agile team. Daily stand up
meeting is a 15 minutes meeting which is held by the team at
the same time everyday during the project. During this meeting,
each team members explain three things which are the progress
of his/her tasks, any problem that comes his/her way, and what
he/she will do today.
4. CONCLUSIONSAgile software development becoming a powerful method to
developing high quality software that meet the customers’
requirements and effectively deal with the problems of rapid
change during the project life-cycle.
The project teams which want to use Agile in their project
must first clearly understand about the characteristics of Agile
projects that significantly different from other “traditional” soft-
ware development methods. Some important characteristics of
Agile projects are blurring roles, continuous development activi-
ties, and team accountability.
Some key factors that can be very useful for improving the
team’s productivity and helping the team to successfully achieve
the goals of the project are co-located team, engaged customers,
self-organizing team, and accountable and empowered team.
3005
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3002–3006, 2015
References and Notes1. A.Alliance,What isAgileSoftwareDevelopment?,AgileAlliance, viewed21Jan-
uary 2015,<http://www.agilealliance.org/the-alliance/what-is-agile/> (2012).2. J. Appelo, Management 3.0: Leading Agile Developers, Developing Agile
Leaders, Pearson Education (2011).3. A. Cockburn, Agile Software Development: The Cooperative Game, Person
Education Inc., Boston (2006).4. A. Manifesto, Manifesto for Agile Software Development, Agile Manifesto,
viewed 21 January 2015, <http://agilemanifesto.org/> (2001).5. D. Leffingwell, Scaling Software Agility: Best Practices for Large Enterprises,
Pearson Education (2007).6. D. Wells, The Values of Extreme Programming, Don Wells, viewed 4 February
2015, <http://www.extremeprogramming.org-/values.html> (2009).7. XProgramming, What is Extreme Programming?, XProgramming, viewed
4 February 2015, <http://xprogramming.com/what-is-extreme-programming/>(2011).
8. D. Wells, Extreme Programming Project, Don Wells, viewed 3 February 2015,<http://www.extremeprogramming.org/map/project-.html> (2009).
9. S. Alliance, Core Scrum, Scrum Alliance, viewed 6 February 2015,<https://www.scrumalliance.org/why-scrum/core-scrum-values-roles>(2014).
10. K. Rubin, Scrum Framework, Agile Atlas, viewed 3 February 2015, <http://agileatlas.org/articles/item/scrum-framework> (2012).
11. T. Dingsøyr, The Journal of Systems and Software 85, 1213,DOI: 10.1016/j.jss.2012.02.033.
12. J. Rasmusson, The Agile Samurai: How Agile Masters Deliver Great SoftwarePragmatic Bookshelf (2010).
13. E. Derby and D. Larsen, Agile Retrospective: Making Good Team Greats, ThePragmatic Bookshelf, Texas (2008).
14. M. Holcombe, Running an Agile Software Development Project, Hogn Wileyand Sons, Inc., New Jersey (2008).
15. J. Shore and S. Warden, The Art of Agile Development, O’reilly Media Inc.,California (2008).
16. A. Miller, How Microsoft’s p&p Teams do Daily Standup Meetings,viewed 2 January 2015,<http://www.ademiller.com/blogs-/tech/2008/07/daily-standup-meetings/> (2008).
Received: 9 November 2014. Accepted: 27 December 2014.
Rudy Yuwono1�∗, Ronanobelta Syakura1�∗, Erni Yudaningtyas1, Endah B. Purnomowati1, and Aisah2�∗
1Department of Electrical Engineering, University of Brawijaya, Malang, 65145, Jawa Timur, Indonesia2Department of Electrical Engineering, Malang, 65145, Jawa Timur, Indonesia
A deployment of the RF devices especially for antenna should be done continuously to obtain the better perfor-mance than the existing. The better performance means that the antenna has the matching impedance closedto perfect condition, higher gain and also better alignment which is related to antenna polarization. To get thebetter performance, the antenna should employ circular polarization which has advantage in alignment. Theoptimization of the antenna should be done to improve the performance. In this paper, the optimized antennacan covers between 2.4 GHz–2.45 GHz of Frequency with the increased gain and circular polarization whichgenerally have better performance compared than existed antenna.
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3007–3009, 2015
Front Rear
Fig. 1. The existed antenna model.8
the frequency of 2.5 GHz as the focus on ISM band. The opti-
mization results are shown at Figures 4–6.
The result of the S11 from Figure 4 explains that the mod-
ified (mod 1 and mod 2) antennas have the advantage than
the existed (et) antenna. For two modified antennas have bet-
ter matching impedance, which indicate from the S11 result least
than −9.54 dB as the tolerance reference of mismatch,7 however
the S11 result only conducted between 2.35 GHz to 2.45 GHz.
From Figure 5 the gain of the modified antennas is increased.
Those are contradictive to existed antenna which has decreasing
trend.
The results from the Figure 6 obtain that the existed antenna
can covers the axial ratio least than 3 dB at each frequency which
indicates that the antenna has circular polarization. It indicates
Fig. 2. Rear view of the modified antenna.
(a) (b)
Fig. 3. Modified antenna models.
Fig. 4. S11 result of the each antenna.
Fig. 5. Gain result.
Fig. 6. The axial ratio result.
different result at modified antennas, which covers 2.35–2.4 GHz
of frequency (mod 1) and covers 2.45 GHz of frequency (mod 2).
4. CONCLUSIONThe antenna improvement results that the modified antenna can
cover 2.4 GHz to 2.45 GHz of frequency. Each antenna has dif-
ferent specification. Both of the antennas have increased gain.
comparing with the existed antenna the two modified antennas
can perform better, however the circular polarization only cover
2.35–2.4 GHz of frequency for first antenna model and 2.45 GHz
of frequency for second antenna model.
References and Notes1. C. A. Balanis, Antenna Theory Analysis and Design, Wiley, USA (2005).2. P. Bhartia, I. Bahl, R. Garg, and Ittipiboon, Microstrip Antenna Design
Handbook, Artech House, USA (2001).
3008
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3007–3009, 2015
3. M. Haneishi and Y. Suzuki, Circular polarization and bandwidth, IEE Electro-magnetic Series 28-Handbook of Microstrip Antennas, edited by J. R. Jamesand P. S Hall, Peter Peregrinus, London (1989).
4. R. Joseph, Studies on Circularly Polarized Broadband Slot Antennas, Ph.D.Thesis, Kumamoto University, http://hdl.handle.net/2298/22943, Japan (2011).
5. R. Joseph and T. Fukusako, Progress in Electromagnetics Research C 26, 205(2012).
6. G. Kumar and K. P. Ray, Broadband Microstrip Antennas, Artech House, USA(2003).
7. S. Makarov, Antenna and EM Modelling with Matlab, Wiley, USA (2002).8. R. Yuwono, R. Syakura, and D. F. Kurniawan, Design of the circularly polar-
ized microstrip antenna as RFID tag for 2.4 GHz of frequency, InternationalConference on Advances Technology in Telecommunication Broadcasting andSatellite 2014, Bali-Indonesia (2014).
Received: 12 September 2014. Accepted: 12 October 2014.
Kamarularifin Abd Jalil1, Nor Shahniza Kamal Bashah1�∗, and Mohd Hariz Naim @ Mohayat2
1Faculty of Computer and Mathematical Sciences, Universiti Teknologi MARA, 40450 Shah Alam, Selangor, Malaysia2Faculty of Information and Communication Technology, Universiti Teknikal Malaysia Melaka, Hang Tuah Jaya,
76100 Durian Tunggal, Melaka, Malaysia
A server is always vulnerable to network attack especially MAC spoofing activity. Current existing MAC spoofingdetection algorithms have several flaws that make protecting the server hard to implement. Thus, a new methodfor detecting MAC spoofing attack is proposed in this research where it is able to detect any spoofing activityand also pinpoint the attacker with minimal false positive.
In order to literally scan for spoofing activity in a network, the
proposed detection method would need to base on the connection
flow as depicted in Figure 2. At first, a genuine initiates HTTP
request connection to the detector/server. Upon connected, the
detector then triggers the ARP request to the connected client to
get the MAC address information. At the same time, the detector
also invoke for UAI that is available via node’s browser. After
has successfully get the MAC address and UAI from the node,
the information is combined together to create a key-like finger-
print that is used later on during authentication. The unique key
is then saved into database and also saves back to the node’s
browser cookies. In the next phase, the attacker will change its
node MAC address to be similar with the legitimate node and
then tries to communicate with the detector/server via HTTP con-
nection. Then, the detector/server trigger again ARP request and
soon discover that duplicated MAC address exist within the net-
work. Now that the duplicated MAC address has been detected,
the detector/server then try to request the key fingerprint that has
been embedded into browser cookie previously. At this stage,
the attacker fails to provide the key fingerprint as similar with
the one in the database and detector can pinpoint that node is
actually a spoofer.
We begin by specifying the notation that will be used.
By running the ARP command, a list of MAC address can
be obtained. The MAC address will then be saved into database
for matching up and authentication. In this research, ARP—a
and ARP—a “Node_ IP_ Address” are used. The ARP—a will
return a list of MAC address and IP address within the same
segment. Meanwhile ARP—a “Node_ IP_ Address” will only
return one MAC address of the specific IP address. For example
running ARP—a 192.168.1.33 will only return one MAC address
of 00-1E-65-C8-2C-6C.
4.4. Setting up Network Test bed
Before MAC address can be captured, the network is setup via
a wireless router or switch. In this test scenario, to imitate a
spoofing attack both node B and C are virtual machine running
Ubuntu 12.04. First, node A is start up and connects to the Access
Point (AP). The IP address is assigned as 192.168.1.34 by the AP.
After has successfully connected, only then the Node C (attacker)
is start up and tried to connect to the AP. This is to mimic that
Fig. 2. Logical flow in detecting spoofing.
3022
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3020–3024, 2015
Fig. 3. Network configuration of node B.
the attacker has already observed and know the MAC address
of node B and thus change his/her node MAC address to be the
same with Node B. The attacker has auto-assigned IP by DHCP
or can assigned IP address manually in case of getting conflict
IP address. In this case, the IP address is manually change by
attacker to 192.168.1.35.
4.5. Changing MAC address of Virtual Machine (VM)
To imitate a spoofing attack, the MAC address needs to be
changed so that there is more than one similar MAC address.
Using the Oracle VM Virtual box, the MAC address of VM can
be modified before starting up. Both of the machines are actually
running from Windows platform at Node A. Then, each of the
VMs network adapter connections are set to Bridge Adapter (BA)
where it is virtually making both machine attached to the AP. In
this scenario, the virtual machine’s MAC address is changed to
00-1E-65-C8-2C-6D similar to Node B.
Figure 3 shows the IP address and MAC address of Node B.
The IP address is 192.168.1.34 while the MAC address is
00-1E-65-C8-2C-6D.
Figure 4 shows the network configuration of the attacker
Node C. Take note that the MAC address has been changed to
be similar with Node B at Figure 3.
4.6. Running ARP Command
Now that the network configurations are all set up and MAC
address had duplicated, the detector Node A should be able to
detect its neighboring nodes. Figure 5 presents the returned MAC
addresses and from here the duplicated spoofed MAC address is
also detected. In order to extract the information so that it can be
Fig. 4. Network configuration of node C.
Fig. 5. List of MAC after running ARP command.
differentiated between legitimate and attacker, some processing is
done at the UAI. The information is retrieved each time the node
make connection to the detector via HTTP request. The informa-
tion are then stored into database for information matching.
4.7. Extracting Information and
Authentication Algorithm
The information retrieved from the ARP command above is then
processed through an algorithm as illustrated in Figure 6. Upon
connecting to the detector or server, the node will be perform-
ing some checking process for authentication purpose. First, the
detector at the server will check whether the MAC address is
registered or not. Initially, if the node has not yet makes any con-
nection, the node MAC address and UAI is registered into the
database. The UAI is also embedded within the node browser.
On the other hand, if the MAC is already registered, the detector
then continues to make inspection whether there is any dupli-
cated MAC detected. If any duplicated MAC is detected, then
the detector request for UAI that was embedded during the ini-
tial connection. The information then will be compared with the
Node Connectto Detector
User Agentinformation is
planted within node
Register MACaddress
MAC address
Is MACregistered?
Isduplicated MAC
Detected?
Detector requestUser Agentinformation
Is informationmatch withdatabase?
Alert/Block Nodefrom Accessing
Server
No
Yes
Yes
No
Yes
Node Informationand Time Connected
Fig. 6. Algorithm overview of spoofing detection.
3023
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3020–3024, 2015
information from the database. If the node fails to supply with
the correct UAI, the node is categorized as a spoofer and then
can be straight away blocked from accessing the server. The node
information such as the IP address and time of connection is also
stored in the database for logging purposes.
Once the system detected that Node C is using similar MAC
address as Node B, a warning message will appear on the screen.
It indicates that the detector senses the duplicated MAC address
value and try request for UAI. Now that the Node C did not have
the same UAI as stored in the database, thus it will get warn-
ing message and the service can be blocked from accessing the
server. The information will also be recorded into the database
for logging purposes.
5. CONCLUSIONSIt can be concluded that detecting MAC spoofing and determin-
ing which one is the attacker is possible via software attestation.
The proposed technique is able to capture MAC address and pro-
cess the information for authenticating and validating connected
nodes. Thus by implementing this technique, any server that pro-
vides services within local network can be protected from MAC
spoofing attack. Network administrator can also be notified in
case of spoofing detected and able to determine which IP address
is performing spoofing attack. In the forthcoming, the proposed
technique of detecting MAC spoofing will be enhance to cope
with its limitation and constraints especially to cater the problem
of different browser UA at client side. It is also hope that the
technique can be used in different network segments particularly
over the Internet for making automatic authentication via MAC
address. Lastly, the problem with MAC spoofing can be solved
or reduce with this new implementation technique.
Acknowledgments: The authors would like to thank Uni-
versiti Teknologi MARA and Malaysian Ministry of Higher Edu-
cation for funding this research under the Fundamental Research
Grant Scheme.
References and Notes1. P. Chumchu, T. Saelim, and C. Srikauy, A new MAC address spoofing detecion
algorithm using PLCP header, IEEE ICOIN (2011).2. Y. Sheng, K. Tan, G. Chen, D. Kotx, and A. Campbell, Detecting 802.11 MAC
layer spoofing using received signal strength, INFOCOM (2008).3. F. Guo and T. Chiueh, Sequence number-based MAC address spoof detection,
8th International Symposium (2006).4. R. A. Redner and H. F. Walker, SIAM Review 26, 195 (1984).5. J. Bellardo and S. Savage, 802.11 denial-of-service attacks: Real vulnerabili-
ties and practical solutions, Proceedings of the USENIX Security Symposium,Washington, D.C., August (2003), pp. 15–28.
Received: 7 November 2014. Accepted: 27 December 2014.
1Department of Computer Science, Kulliyyah of Information and Communication Technology, IIUM2Faculty of Medical and Bioscience, University Technology Malaysia, UTM
Driver behaviour is indeed reckoned to be one of the highest factors affecting fatal accidents. However, majorityof the cases can be avoided if the driver can remain focus and make a correct decision in controlling thevehicle while driving. Decision-making ability of the driver is impeded due to driver behaviour which may involveprecursor emotion of the driver that could lead to fatal accident. Thus, understanding and analyzing the driverbehaviour and the resulting emotion can help prevent accident and reducing accident fatality rate. In this paper,the understanding of precursor emotion of driver is studied in details. This correlation between precursor emotionand their respective emotion can be analysed based on the 2-D Affective Space Model (ASM) using four basicemotions (happy, calm, fear and sad) as stimuli. In this case, the Electroencephalogram (EEG) device is usedto extract brain waves signal while the driver is driving the simulator. The EEG signals are captured through thescalp of the driver and features is extracted using Mel Frequency Cepstral Coefficient (MFCC). Neural networkclassifier of Multilayer Perceptron (MLP) is used to classify the valence and arousal axes for the ASM. Analysisof the precursor emotion for driver shows an interesting finding that complements the discrete classification.In addition, the analysis also indicates how precursor emotion can affect driver behaviour. Consequently, theunderstanding of pre-cursor emotion and its relationship towards driver behaviour could help the driver to controlhis/her emotions while driving which can prevent to fatal accident.
3. T. Bellet, B. Bailly-Asuni, P. Mayenobe, and A. Banet, Safety Science, Elsevier47, 1205 (2009).
4. N. Kamaruddin and A. Wahab, Driver behaviour analysis through speechemotion understanding, Proceeding of the 2010 IEEE Intelligent Vehicle Sym-posium (IV 2010) (2010), pp. 238–243.
5. A. Mehrabian and J. A. Russell, An approach to environmental psychology,MIT Press, Cambridge, MA, USA (1874).
6. P. J. Lang, The three system approach to emotion, edited by N. Birbaumer,A. Ohman, and The Organization of Emotion, Hogrefe-Huber, Toronto (1993),pp. 18–30.
7. P. Jahankhani, V. Kodogiannis, and K. Revett, IEEE Transaction on ModernComputing JVA’06, 120 (2006).
8. A. Savran, K. Ciftci, G. Chanel, J. C. Mota, H. V. Luong, B. Sankur, L. Akarun,A. Caplier, and M. Rombaut, Emotion Detection in the Loop from Brain Sig-nals and Facial Images, Workshop on Multimodal Interfaces, eNTERFACE’06(2006), pp. 11–101.
9. L. A. Sroufe, Emotional Development: The Organization of Emotional Life inthe Early Years, Cambridge Studies in Social and Emotional Development,Cambridge University Press (1997), pp. 58–64.
10. A. S. AlMejrad, European Journal of Scientific Research 44, 640 (2010).11. G. Chanel, J. Kronegg, D. Granjean, and T. Pun, Emotion assess-
ment: Arousal evaluation using EEG’s and peripheral physiological signals,Proceedings International Workshop on Multimedia Content Representation,Classification and Security, Istanbul (2006), pp. 530–537.
12. E. R. Kandel, J. H. Schwartz, and T. M. Jessell, Principles of Neural Science,McGraw Hill (2000).
13. N. M. Nor, A. Wahab, N. Kamaruddin, and H. Majid, Post accident analysis ofdriver affection, 15th IEEE Symposium on Consumer Electronics, ISCE2011,Singapore (2011).
Received: 17 December 2014. Accepted: 8 February 2015.
Pitak Keawbunsong∗, Pitchaya Supannakoon, and Sathaporn Promwong
Faculty of Engineering (Telecommunication Engineering), King’s Mongkut’s Institute of Technology Ladkrabang,Chalongkrung Rd., Bangkok, 10520, Thailand
This article presents the optimization of a Walficsh-Bertoni path loss model to be used in designing the DTTVpropagation in an urban area of southern Thailand through the data collection on the signal power that thenetwork operators of 4 channels broadcasting within the distance of 2.5–6.5 Km. in urban Hat Yai, SongklaProvince, an area of high density with buildings. A least square method is used for the optimization while anefficient indicator is through statistical values of root mean square error (RMSE) and relative error (RE). Theresult is a new Walficsh-Bertoni model that has RMSE value lesser than the original model whereas the REvalue is also closer to zero than the original models. Subsequently, the Walficsh-Bertoni model is more precisein a prediction, making it optimized for use in planning the network.
Keywords: DTTV Propagation, Walficsh-Bertoni Model, Least Square Method.
1. INTRODUCTIONThailand launched the broadcasting of DTTV on April 01, 2014
with DVB-T2 standard and is currently under a process of
installing a main and a sub stations1 and simultaneously plans on
a design of a gap filler station in an urban area that faces with a
dead spot caused from the environment such as high rise build-
ings and their density that obstruct the TV signal.2 An effective
planning for the gap filler that includes the ability to cover the
proper areas, the use of a suitable transmitted power and the cost
saving requires an accurate predicted path loss model for the area.
Obviously, the development of the predicted model for the areas
using a least square method has been studied and presented in
many researches.3�4
Walficsh-Bertoni path loss model was initiated on the basis
of the loss from the diffraction of waves in the high rise areas
as well as the study on the signal receiving between buildings
at the diffraction of the waves in various directions.5 Thus, this
becomes interesting to apply for use in a design of a network
gap filler station within an urban area.
This article presents an optimized Walficsh-Bertoni path loss
model in urban areas of southern Thailand through using a least
square method. The second part illustrates a Walficsh-Bertoni
model in details whereas the third explains the data collection
for the optimization while the forth demonstrates an optimization
∗Author to whom correspondence should be addressed.
process of a path loss model and the fifth explains the result,
finally, the sixth is the conclusion.
2. WALFICSH-BERTONI MODELWalficsh and Bertoni present a semi-empirical path loss model6
for use in the environment of high rises and density as the model
assumes that the height of the buildings is a function of uniform
distribution. If the distance between buildings is equal, the wave
propagation will turn in various directions, through the building
rows onto the received antenna, as Figure 1.
A Walficsh-Bertoni model consists of 3 parts of path loss: free
space loss (PLfs), diffraction loss from rooftops (PLrooftops) and
diffraction and scatter loss from rooftops down and from a street
(PLdown�7
PLfs =−10 log��/4�r�2 (1)
PLdown = ��1
2�2�HB −hm�(2)
PLrooftops =[
0�1
(sin
√d/�
0�03
)0�9]2
(3)
sin can be written in terms of a height of a transmitted
antenna �ht�, a building height (HB), a distance (R) and the equa-
The least square method8 is a function of the sum when deviation
square becomes minimum.
P�a�b� c� � � ��=n∑
i=1
�yi−ER�xi�a�b� c� � � ���2 = min (14)
yi = value obtained from the measurement at the distance of xi.ER�xi�a�b� c� � � ��= the result from the predicted model at the
distance of xi.a, b, c = model parameter for the optimization.
n= the number of data from the measurement.
All partial differential results of P function should be equal to
zeros
�P/�a= 0� �P/�b = 0� �P/�c = 0 (15)
From (11)–(13) we can spread them to:
a= E0 +Esys� b = �sys (16)
When logR= x from (16), a new equation can be written as:
ER = a+bx (17)
From (17), we can look for a constant value of a, b from the
measured data group and the (17) can be spread into:
n∑i=1
(�yi−ER�xi�a�b�c��
�ER
�a
)=∑
�yi−a−bxi�·1=0 (18)
n∑i=1
(�yi−ER�xi�a�b�c��
�ER
�b
)=∑
�yi−a−bxi�·xi=0 (19)
2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 780
90
100
110
120
130
140
150
160
Distance (Km)
Pat
h lo
ss (
dB)
CH26 (fc = 514 MHz)
Measurement path lossWalficsh-Bertoni modelOptimized Walficsh-Bertoni model
Fig. 3. The comparison of pre and post optimization of Walficsh-Bertonimodel at CH26 with the frequency of 514 MHz.
2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 780
90
100
110
120
130
140
150
160
Distance (Km)
Pat
h lo
ss (
dB)
CH42 (fc = 642 MHz)
Measurement path lossWalficsh-Bertoni modelOptimized Walficsh-Bertoni model
Fig. 4. The comparison of pre and post optimization of Walficsh-Bertonimodel at CH42 with the frequency of 642 MHz.
(18) and (19) can be re-spread into:
n ·a+b∑
xi =∑
yi� a∑
xi +b∑
x2i =
∑�xiyi� (20)
Replace the variable a�b in (20) and we get a statistical estimated
value of the parameter a�b as:
a=∑x2i
∑yi−∑
xi∑xiyi
n∑x2i −�
∑xi�
2� b= n
∑xiyi−∑
xi∑yi
n∑x2i −�
∑xi�
2(21)
From (21), we can get the value a�b of the path loss data group
from the measurement and from (7) and (16), we get the value
of the offset and slop of the original Walficsh-Bertoni path loss
model, as:
E0new = a−Esys� �sysnew = b
38(22)
5. FINDINGS ANALYSISThe optimization of the Walficsh-Bertoni model to measure the
efficiency is done by looking for a statistic value of RMSE and
2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 780
90
100
110
120
130
140
150
160
Distance (Km)
Pat
h lo
ss (
dB)
CH46 (fc = 674 MHz)
Measurement path lossWalficsh-Bertoni modelOptimized Walficsh-Bertoni model
Fig. 5. The comparison of pre and post optimization of Walficsh-Bertonimodel at CH46 with the frequency of 674 MHz.
3032
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3030–3033, 2015
2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 780
90
100
110
120
130
140
150
160
Distance (Km)
Pat
h lo
ss (
dB)
CH54 (fc = 738 MHz)
Measurement path lossWalficsh-Bertoni modelOptimized Walficsh-Bertoni model
Fig. 6. The comparison of pre and post optimization of Walficsh-Bertonimodel at CH54 with the frequency of 738 MHz.
RE from the measured signal power value which can be found
from the following equations.9
RMSE =√∑
�PLm −PLapprox�2
N −1(23)
RE = �PLm −PLapprox��PLm�
(24)
PLm is the measured path loss value; PLapprox is the Walficsh-
Bertoni path loss value; N is the amount of measured data where
the amount of 2300 points of each channel is selected for this
research.
The optimized result of the Walficsh-Bertoni model is the con-
stant parameter a�b, E0new and �sysnew as shown in Table II.
It is noticeable that the value of E0new decreases when the fre-
quency increases whereas �sysnew increases when the frequency
increases. The statistical comparison shown in Table III indicates
that the new Walficsh-Bertoni model is more accurate than the
original one. This is because, after the optimization, the values
of RMSE and RE reduce and they reduce when the frequency
increases. Figures 3–6 compare the pre and post optimization of
the frequency of each channel. It can be observed from the curve
line that in Figure 3; CH26; 514 MHz, the post optimization of
Walficsh-Bertoni path loss model is slightly close to the mea-
sured data whereas in Figure 4; CH42; 642 MHz, the curve line is
closer to the measured data than the one of Figure 3. In Figure 5;
CH46; 674 MHz, the curve line is closer to the measured data
than the one of Figure 4 and in Figure 6; CH54; 738 MHz, the
curve line is again closer to the measured data than the one of
Figure 5. This is consistent with the statistic result showing that
the new model are closer to the measured data that are higher
than the original one and that the new Walficsh-Bertoni model of
the high frequency is closer to the measured data than the ones
of low frequency.
6. CONCLUSIONThe result of Walficsh-Bertoni model obtained from the opti-
mization shows the suitability for use in designing a gap filler
station of the DTTV propagation in the urban Hat Yai, Songkla
Province where the dead spot exists. The model can also be
used as a functional reference for the urban areas in the south
of Thailand in order to appropriately cover the areas. The abil-
ity to use the right broadcasting power leads to cost reduc-
tion. The statistical efficiency indicator of each signal frequency
shown in Table IV reveals the decreasing value of RMSE and RE
upon being compared with the original models prior to the opti-
mization as the value decreases when the frequency increases.
Figures 3–6 demonstrate a clear result in which the Walficsh-
Bertoni model is closer to the measured data than the origi-
nal models. The research findings on Walficsh-Bertoni path loss
model optimization demonstrate the frequency variations that are
closer to the measured data. Seeking for the optimized upper
and lower boundaries of the path loss model through a confident
interval estimation method in order to create the more precise
prediction can be a significant future study topic.
References and Notes1. Office of The National Broadcasting and Telecommunication Commission, The
Radio Frequency for Digital Terrestrial Television in Thailand, NBTC, December(2012).
2. S. R. M. de Carvalho, Y. Iano, and Rangel, ISDB-TB field trials and cover-age measurements with gap-filler in suburban environments, IEEE InternationalSymposium on Broadband Multimedia Systems and Broadcasting, Nuremberg,Germany, June (2011).
3. J. Chebil, A. K. Lwas, Md. R. Islam, and Al-Hareth Zyoud, Proc. CSIT 5, 252(2011).
4. A. Bhuvaneshwari, R. Hemalatha, and T. Satyasavithri, Statistical tuning ofthe best suited prediction model for measurements made in Hyderabad cityof Southern India, Proceedings of the World Congress on Engineering andComputer Science, San Francisco, USA, October (2013).
5. I. Joseph and B. Michael, American Journal of Physics and Applications 30, 10(2013).
6. J. Walfisch and H. L. Bertoni, IEEE Transection on Antenna and Propagation36, 1788 (1998).
7. J. Isabona and S. Azi, International Journal of Engineering and InnovativeTechnology 2, 14 (2012).
8. B. S. L. Castro, M. R. Pinheiro, and G. P. S. Cavalcante, Optoeletronics andElectromagnetic Applications 10, 106 (2011).
9. R. Mardeni and K. F. Kwan, Progress in Electromagnetics Research 13, 91(2010).
Received: 30 September 2014. Accepted: 3 November 2014.
Angela Amphawan1�2�∗ , Sushank Chaudhary1, and Tse-Kian Neo3
1Integrated Optics Group, School of Computing, Universiti Utara Malaysia, Sintok, Kedah, Malaysia2Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge, Massachusetts, 02139, USA
3Faculty of Creative Multimedia, Multimedia University, Cyberjaya, Selangor, 63100, Malaysia
In keeping pace with the exponential growth in network traffic, it is imperative to explore new multiplexing degreesof freedom in addition to wavelength, time, intensity and phase. The next frontier in free-space optics (FSO)envisages the eigenmode of an optical resonator as a new multiplexing degree of freedom. This paper focuseson mode division multiplexing of two parallel 2.5 Gbps channels on spiral-phased Hermite-Gaussian modesHG 10 and HG 11 for free-space optical interconnects at a wavelength of 850 nm. The signal-to-noise ratios,eye diagrams and modal analyses are investigated.
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3051–3054, 2015
Fig. 1. MDM of spiral-phased HG 10 and HG 11 modes for indoor FSO link.
2. SYSTEM DESCRIPTIONFigure 1 shows a block diagram of proposed FSO model using
spiral-phased HG modes designed in OptiSystem software.22 As
shown in Figure 1, the proposed architecture consists of two
independent non-return-to-zero (NRZ) encoded channels, each
carrying 2.5 Gbps data stream over an 850 nm optical spatial
carrier. Two laser modes of spot size 5 �m are used for data
transmission: HG 1, 0 for Channel 1 and HG 1, 1 for Channel 2.
The HG mode is described mathematically as:23
�m�n�r��� = Hm
(√2x
wo�x
)exp
(− x2
w2o x
)exp
(j�x2
�Rox
)
Hn
(√2y
wo�y
)exp
(− y2
w2o y
)(j�y2
�Roy
)(1)
A vortex lens is used to apply a spiral phase transformation
to each HG mode generated as shown in Figure 2. The applied
phase is given by the following:
T �x�y� = exp
[−�n�x2 +y2�
2�f+m tan−1
(x
y
)](2)
where f is the focal length of the lens, m is the vortex index
and n is the refractive index. For Channel 1, the vortex lens has
vortex index, m= 2 whereas for Channel 2, m= 5. The transverse
electric field used for the two modes used in Channel 1 and
Channel 2 are depicted in Figure 2. The output of two channels
are combined, amplified and transmitted through free-space of
varying distances from 200 m to 1000 m. The link is free from
atmospheric turbulences and suited for indoor applications. The
link equation for free space optics29 is modelled by:
PReceived = PTransmitted
(d2RR
�dT +�R�2
)10−�R/2 (3)
where dR defines receiver aperture diameter, dT is the transmitter
aperture diameter, � is the beam divergence, R is the range and
� is the atmospheric attenuation. At the receiver side, the trans-
mitted mode is extracted based on non-interferometric modal
decomposition.30 The output mode is then fed to a spatial PIN
detector followed by low-pass Gaussian filter to retrieve the orig-
inal baseband signal.
3. RESULTS AND DISCUSSIONIn this section, results from our proposed MDM based FSO link
are presented and discussed. The signal-to-noise ratio (SNR) and
total received power at the receiver are shown in Figures 3 and 4
respectively. Both SNR and total received power graphs demon-
strate that Channel 1 carrying spiral-phased HG 10 performs
better than Channel 2 carrying spiral-phased HG 11. The SNR
deteriorates with FSO range for Channel 1 with SNR values of
34.11 dB, 19.36 dB and 9.19 dB for Channel 1 for a FSO link
of 200 m, 600 m and 1000 m respectively. The SNR for Chan-
nel 2 declines with a slightly steeper slope, with SNR values of
30.52 dB, 10.11 dB and 2.32 dB for a FSO range of 200 m,
(a)
(b)
Fig. 2. Excited modes (a) HG 10 mode with vortex index, m = 1 forchannel 1 (b) HG 11 mode with vortex index, m = 3 for channel 2.
3052
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3051–3054, 2015
Fig. 3. SNR versus FSO range.
600 m and 1000 m respectively. On the other hand, the total
power received for Channel 1 are −65.43 dB m, −80.31 dB m
and −90.14 dB m whereas for Channel 2, the total received
power is −69.22 dB m, −89.18 dB m and −97.32 dB m for a
FSO link of 200 m, 600 m and 1000 m respectively.
The modal content at the receiver is computed in terms of
power coupling coefficient of linearly polarized (LP) modes
using noninterferometric modal decomposition,19 arranged in the
order of descending power coupling coefficient, as shown in
Figure 5. For Channel 1, the power is coupled predominantly into
modes with azimuthal mode order of 2, with the largest power
coupled into mode LP 0,1, followed by LP 2,1; LP 2,2; LP 0,2
and LP 0,3. For Channel 2, the power is coupled predominantly
into modes with an azimuthal mode order of 5, with the largest
power coupled into mode LP 1,1 followed by LP 5,1; LP 5,2;
LP 5,3; LP 5,4 and LP 4,7. This agrees with the spatial profiles
of the spiral-phased HG modes. Wide eye openings are attained
for both channels are shown in Figure 5, which confirm success-
ful data transmission of 2× 2�5 Gbps over a FSO link of 600
meters for Channel 1 and 400 meters with Channel with accept-
able SNR.
Fig. 4. Total received power against FSO range.
(a)
(b)
Fig. 5. Modal decomposition at the receiver in terms of descending orderof power coupling coefficients in linearly polarized modes: (a) Chanel 1(b) Channel 2.
4. CONCLUSIONIn this work, 2 × 2�5 Gbps data transmission is realized for a
400 m indoor FSO link by MDM of two independent channels on
spiral-phased HG 10 and HG 11 modes. The results reveal that
Channel 1 propagating spiral-phased HG 1,1 mode with vortex
index, m = 1 is more robust than Channel 2 propagating spiral-
phased HG 0,1 mode with vortex, m = 3. The model may find
applications in optical interconnects in mega data centers.
References and Notes1. S. Chaudhary and A. Amphawan, Journal of Optical Communications 35, 327
(2013).2. A. K. Majumdar and J. C. Ricklin, Free-Space Laser Communications: Princi-
ples and Advances, Springer, New York (2008), Vol. 2.3. A. M. Khalid, G. Cossu, and E. Ciaramella, Diffuse IR-optical wireless system
demonstration for mobile patient monitoring in hospitals, 2013 15th Interna-tional Conference on Transparent Optical Networks (ICTON) (2013), pp. 1–4.
4. L. Chevalier, S. Sahuguede, and A. Julien-Vergonjanne, Investigation of obsta-cle effect on wireless optical on-body communication performance, 2014 21stInternational Conference on Telecommunications (ICT) (2014), pp. 103–107.
5. Y. Wang, X. Huang, L. Tao, and N. Chi, 1.8-Gb/s WDM visible light com-munication over 50-meter outdoor free space transmission employing CAPmodulation and receiver diversity technology, Optical Fiber CommunicationConference, Los Angeles, California, (2015), p. M2F.2.
6. A. Aladeloba, M. Woolfson, and A. Phillips, J. Opt. Commun. Netw. 5 (2013).
3053
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3051–3054, 2015
7. (2015, Feb. 3, 2015) Cisco Visual Networking Index: Global Mobile Data Traf-fic Forecast Update 2014–2019 White Paper.
8. K. Nisar, A. Amphawan, and S. B. Hassan, Int. Journal of Advanced Pervasiveand Ubiquitous Computing (IJAPUC) 3, 50 (2011).
9. !!! INVALID CITATION !!!10. S. O. Arik, D. Askarov, and J. M. Kahn, J. Lightwave Technol. 32, 1841
(2014).11. S. O. Arik, J. M. Kahn, and K.-P. Ho, IEEE Signal Processing Mag. 31, 25
(2014).12. A. Amphawan, Optics Exp. 19, 23085 (2011).13. A. Amphawan, V. Mishra, and K. N. B. Nedniyom, J. Mod. Opt. 50, 1745
(2012).14. C. P. Tsekrekos and D. Syvridis, IEEE Photon. Technol. Lett. 24, 1638 (2012).15. Y. Jung, R. Chen, R. Ismaeel, G. Brambilla, S. U. Alam, I. P. Giles, et al.,
Optics Express 21, 24326 (2013).
16. A. Amphawan and N. M. A. A. Samman, Tiering effect of solid-core pho-tonic crystal fiber on controlled coupling into multimode fiber, SPIE OpticalEngineering+Applications: Photonic Fiber and Crystal Devices: Advances inMaterials and Innovations in Device Applications VII, San Diego (2013).
17. J. Wang, J.-Y. Yang, I. M. Fazal, N. Ahmed, Y. Yan, H. Huang, et al., Nat.Photon. 6, 488 (2012).
18. Y. Ren, H. Huang, G. Xie, N. Ahmed, Y. Yan, B. I. Erkmen, et al., OpticsLetters 38, 4062 (2013).
19. J. Cang, P. Xiu, and X. Liu, Optics and Laser Technology 54, 35 (2013).20. R. Chen, L. Liu, S. Zhu, G. Wu, F. Wang, and Y. Cai, Optics Express 22, 1871
(2014).21. L. Pan, C. Ding, and H. Wang, Optics Express 22, 11670 (2014).22. Optiwave, Optisystem, edited by Ottawa, Canada (2014).23. J. Enderlein and F. Pampaloni, Journal of the Optical Society of America A
21, 1553 (2004).
Received: 23 January 2015. Accepted: 18 April 2015.
W. Yohana Dewi Lulu1�2�∗, Adhistya Erna Permanasari1, Ridi Ferdiana1, and Lukito Edi Nugroho1
1Electrical Engineering and Information Technology, Gadjah Mada University, Yogyakarta 55281, Indonesia2Information System, Polytechnic Caltex Riau, Pekanbaru 28265, Indonesia
The organizations use data as an important asset. Data became the important assets in some organizationsmay give effect due to the cost. Data quality standards for non-profit-oriented education organizations have notspecifically defined. This study aims to provide guidance for non-profit organizations in order to produce qualitydata. The focus of analysis is a data modeling, determination of dimensions of quality and maturity level ofdata quality in organizations. The first step needs to be done is to create a data quality model based on entityrelationship data and to fix dimensions of data quality to be used. Dimensions are used to follow the intrinsic,contextual, representative, accessibility data quality. Determination of the level of maturity of the initial and finalof data quality analysis process used to make recommendations next stage. The results showed that stages ofdefining quality varied by the user complicate the process of preparing the data model quality. Maturity of dataquality management were mostly at the defined data management (level 2) and the process move to level 3.Maturity data quality management requires analysis related to data and process. Evolving business processorganization should be aligned with the data quality standards have been set.
Keywords: Data Quality, Maturity of Data Quality, Higher Education.
1. INTRODUCTIONData are very useful in an organization. Data can be an asset to
the organization. An understanding of the functions and bene-
fits of the data in general, should be shared by all parts of the
organization. The data would be useful if data are presented in
accordance to the needs of their owners. Usefulness of data varies
according to the needs of their owners. This causes differences
in the benefits of data for every data owner.
The use of data in many organizations is used to basis for
decision-making to carry out operational and strategic level.1–3
Data quality is a critical factor for achieving strategic and oper-
ational business.4 Poor data quality will have a negative impact
in many ways in organizations.5 If the data is of poor quality
but it did not be identified and corrected, it could have economic
and social impacts significantly negatively on an organization.
Efforts to improve the quality of data for a particular purpose or
long-term goals require improvements in the quality of data that
involves a data-driven and process-driven.5–7
Knowledge about data’s purpose right from the beginning of
process can be contribute to increase data quality.8 There is one
perspective that can be used to analyze and improve data quality.5
This paper provide a comprehensive process for exploration data
∗Author to whom correspondence should be addressed.
quality maturity. The purpose of this paper is to describe an
understanding of the quality of data within the organization with
the concept of maturity. This research was conducted by study-
ing previous research literature and analysis of case studies at a
small higher education.
The result show benefit and limitations of the approach, allow-
ing practitioner to tailor the approach to their needs. Analysis
of process driven that used in business processes for marketing
process in the case study, demonstrating the benefits, a sense of
ownership and a better understanding of the quality of data.
The remainder of the paper is structured as follow Section 2
introduces the basic theory of data quality and maturity and also
basic theory of business process. Section 3 analyse case study,
Section 4 report the result and finally Section 5 presents the
Conclusion.
2. BASIC THEORY OF DATA QUALITY,
MATURITY AND BUSINESS PROCESSThe main rationale underlying this paper is to achieve an under-
standing of the quality of data within the organization with the
concept of maturity. In this section we will explain the concept
of data quality, business process management and business pro-
cess management maturity, based on the study of literature that
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3060–3064, 2015
2.1. Data Quality
The quality of the existing data is very important for an orga-
nization. The definition of quality varies greatly various organi-
zations and academia. The International Standard Organization
(ISO) defines quality as degree to which a set of inherent char-
acteristics fulfils requirements. A collection of data is of high
quality, in customers view, if the data meet his, her or its needs.9
According to Genichi Taguchi,10 quality is closeness to target and
deviations means loss to society. However, J. Juran and American
Society of Quality Control referred quality as fitness for use11
and fitness is defined by customers.12–14 Organization study qual-
ity of data based on the definition of fitness for use and fitness
is customers satisfaction for their purposes.
Data quality factors can be divided into three domain: data
service, data value, data structure quality. Figure 1 illustrates a
recurring stage of evaluating the quality of the data from the
three elements (structure, value and service). Management of data
quality includes the level of depth and greater use of quality
components.
The greater use of data quality covering the integration of
information systems within the scope of the organization’s
database. Greater use of the data quality including quality lev-
els defined data structures and process integration as well as the
overall consistency of the data.15 However, it is hard to evalu-
ate data quality objectively without considering the width of data
under the discrete and complicated information system enviro-
ment data quality can be assessed from the standpoint of the
product. Therefore, the data has producer, consumers and brokers
in the process of fulfilling the data itself.14
Data is a collaboration of two components, namely the data
model and the data value.16 The data model describes the con-
dition of the real world with structured. Entity data model
expressed as objects, attributes and ideas in the real world.
Attributes in the data model involves the nature of entities and
relationships between entities. Value of data provides information
realization of an attribute in the entity or association. The concept
of data quality is the same thing with the quality of the product.
The quality of the product produced depends on the design and
production process itself. Understanding the meaning of quality
and how quality is measured will assist in the design process of
good quality. The quality of data that involves many factors and
components referred to as a multidimensional concept.17�18
2.2. Data Quality Assesment
Data quality assessment can provide an idea of the level of qual-
ity of the existing data and establish appropriate criteria for data
quality improvement process further. In their improvement steps,
methodologies adopt two general types of strategies, namely
data-driven and process-driven.6�19 Data-driven strategy can help
the process of improving the quality of data with making changes
and updates to the value of the data. Process-driven strategies can
help to improve the quality of data to make changes and modify
the data creation process.2�6 As an example, a process can be
Fig. 1. Data quality model.15
redesigned by including an activity that controls the format of
data before storage.
Process-driven data quality management aims to identify root
cause of poor data quaility. Process-driven conducts process con-
trol or redesign activities. In process driven, process modeling is
providing the means to understood and commicate.6 Data gener-
ated in the right business processes that are expected to have the
right quality. Many organizations often forget this, organizations
pay more attention to the data resulting from the process of how
to get it.
Data quality methodology apply the concept of data-driven
and process-driven as one of the main strategies in activities
to improve the quality. Simply put, a data-driven strategy can
improve data quality by directly modifying the data values, and
strategy driven process of improving quality by redesigning pro-
cesses that will create or modify the process of inputting the
data and data processing. Some methodologies use strategies,
phases, activities and different quality dimensions, but the stages
of assessment and improvement has always been a part that is
not abandoned.
For the assessment phase, the diagnostic process data quality
and relevant quality dimensions, using adequate data quality tools
(DQT). The increase mainly focus:
(a) The process of identifying the root of the error.
(b) Make improvements of error with the right DQT and
(c) Make changes to the design of specific DQ engineering and
redesign of data creation process to improve its quality.
Batini2 presents and compares some of the most widesread
methodologies.
2.3. Data Quality Management Maturity
Data quality evaluation and management should be defined
from many points of view such as total corporate integration
management,20–22 data structure quality management, and man-
agement maturity stages. The data management maturity model
that is used as the capability maturity model of software pro-
cess evaluation (Fig. 2).15�22�23 According to Ryu,15 he proposed
maturity stage that consists of four levels, namely initial, defined,
managed and optimized.
Level 1. Initial (Data Management Stage). This stage defines the
structure of the quality of the data and rules used in the database
by management.
Level 2. Defined data management. This stage deals with mod-
eling of data models, both logically and physically. This phase
establishes the model data according to pre-established database.
Level 3. Managed. These stages are managed using standard
enterprise. Management of metadata at this stage includes the
entire enterprise data. The integration process is developed at this
stage.
Level 4. Optimized. Stages set management in the form of archi-
tectural models. This stage optimizes data management, data
models and relationships within the company standards.
2.4. Basic Theory Business Process
The process is a stage arranged in a logical work in creat-
ing goods and services, including the creation of the use of
resources.24�25 Process-driven data quality management is used
to find the root cause of data quality is low.20�26 In contrast with
data correction, process defects repaired and adjusted to main-
tain improvement. Modeling process should provide a means to
3061
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3060–3064, 2015
Fig. 2. Data quality management maturity model.15
understand and communicate the process to be able to control
and redesign activities.2 Quality management based data-driven
process is not limited in the domain of information systems.
Process-driven data quality needs to be supported by other factors
that may be related, the definition of data quality and suitability
of the dimensions used.
Concerning process-driven activities, a typical business pro-
cess reengineering activity,27 for a comprehensive discussion is
composed of three activities: Mapping and analyzing the as-is
process; Designing the to be process, according to one or multi-
ple alternatives; Implementing the reengineered process and con-
tinuously improve. Business processes become key elements of
success of an organization.
Business process describes how the organization operated and
how the role of the organization’s performance. Organizations
need to strive to get better performance. Therefore organizations
need to understand the maturity of their business processes.28�29
Business Process Management (BPM) is concerned with the effi-
cient management of business processes and their continuous
improvement.30 For managing processes, BPM provides a set of
structured methods and technologies.31 Current research provides
an overview of the several BPM standards across the BPM life
cycle.32 Several BPM maturity models exist to assess organiza-
tions maturity and provide guidance for its improvement.
3. ANALYSIS OF CASEPolytechnic Caltex Riau (PCR) was built by PT. Caltex Pacific
Indonesia (CPI), BP Migas and the Government of Riau
Province, as proof of their commitment to improving the quality
of human resources in Riau. PCR aims to produce quality human
resources in the field of applied technology, which has the knowl-
edge and integrity. The resulting human resources are expected
to meet the manpower needs in the industry both nationally and
internationally.
PCR is trying to focus on the process of his own efforts to
achieve the vision and mission. One of the very early process
of note is the new admissions process. This process is the input
to the production process. Since four years ago, new admis-
sions process had to use information technology as a marketing
medium. Then, marketing team began to realize the role of the
data obtained in the previous year is very useful for the next
admissions process.
PCR required to be more competitive advantage to the other.
Good data management will greatly assist the higher education in
many ways. Data quality in higher education need to be examined
more carefully to support various processes in higher education,
such as the process of marketing, learning data processing and
other learning outcomes.
At this time PCR as one of higher education that focuses on
the skills of its graduates have more than 10 years old. This
year the number of active students who are studying in the PCR
reached 1600 student. Since its inception, PCR-based managed
IT concepts. Processing of the data generated by the share in the
PCR process has been managed IT, although not achieving the
integration process.
Maturity level assessment of data quality management will be
done by providing a preliminary question. Assessment given to
the users of the application for 60 user PCR. The results of the
assessment have been distributed only returned 39 pieces. The
data can be validated and accepted in the processing of as many
as 37.
Based on the results in Figure 3 Level satisfied data quality
management shows that the management of the quality of exist-
ing data has reached 73% level of satisfaction. This result does
not give a picture directly to the successful management of data
quality.
Therefore conducted direct interviews with five users are lim-
ited by user applications with the same end goal. Based on the
results of interviews conducted, showed that the level of satis-
faction about the quality of the data that is otherwise still based
on the viewpoint of the work that has been defined for them.
The absence of standards managed organization as a whole in
the process of data quality.
Data quality dimensions assessed include four aspects of data
quality. Intrinsic aspect of assessing the dimension accuracy and
believability. Contextual aspects of the quality of the data that
already includes the process of software engineering require-
ments only assess the dimensions of completeness. It is based
that the data provide the best quality if according to its users.
In the aspect of data quality assessed dimensional representation
of consistency and duplicate, whereas the accessibility aspects of
the assessment on both dimensions.
Determination of dimensions which are valued at the ini-
tial survey and interview data users within the scope of the
Fig. 3. Level satisfied data quality management.
3062
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3060–3064, 2015
application. Survey and initial interviews covering all aspects of
data quality version of Wang.18 Assessment is done by using
the interval 1–9 to provide flexibility assessment. Results of the
importance of data quality dimensions are given in Figure 4 Level
Important Data Quality Dimension.
In 4 Level Important Data Quality Dimension it can be seen
that the component dimensions have the highest level of interest
is the believability and then the dimensions of access security.
Level of understanding of leaders and managers of higher educa-
tion on the concept of quality management data is done with the
interview. It is the focus of the interview, relating to the design
process associated with the data, data storage and data retrieval.
Components of data quality are an indicator for each data collec-
tion process. Data collection was done through interviews with
the leaders of the organization.
Clarification of data quality aspects define earlier, clarified and
adapted to the existing file governance, management, marketing
staff and IT staff. Overall respondents were given a brief under-
standing of the concepts and data quality management maturity.
This is used to provide the same level of understanding of the
quality of the data.
To date, this institusion do not have a company data guide-
lines nor a management data center initiative. For example, they
still use the same attribute with different modifications. They had
difficulty in defining the marketing data and perform risk man-
agement and control it manually.
Data used and created in times of need in the process. Associ-
ated with the process used to identify the data used, the necessary
Awareness, Exploration, Reporting, Fixing and Preventing. Man-
agement believes that root cause analysis is required from the
beginning until today to be able to fix the attribute data needed
in the marketing process.33
4. RESULTThe main driver for PCR to undertake the data quality manage-
ment initiative was costumer data, which is in line with the most
relevant areas for increase the marketing process and to maintain
the customer. By applying the data quality framework. PCR can
have began bottom-up, being only focused on a particular dataset
and not having a organization data policy, strategy or governance.
Utility driven approach can be used to better focus on specific
database.34
Results of analysis and interviews conducted on PCR showed
that PCR is still at the stage of defining data management
Fig. 4. Level important data quality dimension.
(level 2). This is evident because they are still improving business
processes and choosing new tools for data management. The next
stage they are going is the stage to prevent data quality problems
and must make centralized management. Management standards,
rules and classifications for all business processes.
The concept of intrinsic DQ gives meaning that the data has
a quality of its own point of view. DQ contextual concept is
discussed that the data quality requirements must be taken into
consideration in the process of how the data is obtained and
used. Representation and accessibility DQ how do system focuses
more on the quality of data. All of this concept provides an
understanding that high quality data must be intrinsically good,
contextually appropriate to the task, clearly represented, and can
be accessed by the consumer data.
5. CONCLUSIONSData quality management maturity begins in the standard-setting
process the desired data quality user. Maturity management data
quality requires analysis of the relationship between the data and
processes. Developments in the organization’s business processes
must be aligned with the data quality standards that have been
set. This also applies vice versa, where the standard of data qual-
ity can be increased or changes with the organization’s business
processes.
In our case study in PCR, the level of maturity of the quality
of the data is still in the understanding of data quality standards.
If no explanation at the beginning of the process of data qual-
ity, better understanding of data quality processes that already
exist and of all the different users. Further development can be
done by involving more components of organizational business
processes to gain a broader picture and depth data quality man-
agement maturity. Further assessment of the various viewpoints,
such as the application point of view, and private organizations
or user standpoint, would be required. Maturity assessment needs
to consider elements of the user. How, who and what the role of
the user in the process of an organization’s data quality. How and
to what extent a user to understand the level of need for quality
data.
References and Notes1. M. Eppler and M. Helfert, A classification and analysis of data quality costs,
International Conference on Information Quality (2004).2. C. Batini, C. Cappiello, C. Francalanci, and A. Maurino, ACM Computing Sur-
veys CSUR 41, 16 (2009).3. A. Even and G. Shankaranarayanan, Journal of Computer Information Sys-
tems 50 (2009).4. M. H. Ofner, B. Otto, and H. Österle, Business Process Management Journal
18, 1036 (2012).5. A. Haug, F. Zachariassen, and D. V. Liempd, Journal of Industrial Engineering
and Management 4, 168 (2011).6. P. Glowalla, P. Balazy, D. Basten, and A. Sunyaev, Process-driven data quality
management—An application of the combined conceptual life cycle model,Hawaii Internasional Conference on System Science (2014), Vol. 47, p. 9.
7. P. Glowalla and A. Sunyaev, Managing Data Quality with ERP Systems-Insights from the Insurance Sector (2013).
8. Y. W. Lee and D. M. Strong, Journal of Management Information Systems20, 13 (2003).
9. T. C. Redman, Data Quality Management Past, Present, and Future: Towardsa Management System for Data: Handbook of Data Quality Research andPractice, Springer, Berlin, Heidelberg (2013).
10. G. Taguchi, Introduction to Quality Engineering: Designing quality into Prod-ucts and Processes (1986).
11. J. M. Juran, Juran on Leadership for Quality, Simon and Schuster (2003).12. T.-V. Tran, S. Kim, and M.-H. Hsiao, Data Stewardship and Flow Management
for Data Quality Improvement.
3063
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3060–3064, 2015
13. M. F. Bosu and S. G. MacDonell, A taxonomy of data quality challenges inempirical software engineering, 22nd Australian Software Engineering Con-ference ASWEC, 2013 (2013), pp. 97–106.
14. L. Sebastian-Coleman, Measuring Data Quality For Ongoing Improvement:A Data Quality Assessment Framework, Newnes (2012).
15. K. S. Ryu, J. S. Park, and J. H. Park, ETRI Journal 28, 191 (2006).16. T. C. Redman, Digital Press (2001).17. R. Y. Wang and D. M. Strong, Journal of Management Information Systems
5, (1996).18. Y. Wand and R. Y. Wang, Communications of the ACM 39, 86 (1996).19. L. L. Pipino, Y. W. Lee, and R. Y. Wang, Communications of the ACM 45, 211
(2002).20. T. C. Redman and A. Blanton, Artech House Inc. (1997).21. G. Shankaranarayan, M. Ziad, and R. Y. Wang, Journal of Database Manage-
ment 14, 14 (2003).22. J. Herbsleb, D. Zubrow, D. Goldenson, W. Hayes, and M. Paulk, Communica-
tions of the ACM 40, 30 (1997).23. J. Kaur, Comparative Study of Capability Maturity Model (2014).24. H. U. Buhl, M. Röglinger, S. Stöckl, and K. S. Braunwarth, Business and
Information Systems Engineering 3, 163 (2011).25. J. V. Brocke and M. Rosemann, Handbook on Business Process
Management 2, Springer (2010).
26. L. P. English, Improving Data Warehouse and Business Information Quality,John Wiley & Sons (1999).
27. S. Muthu, L. Whitman, and S. H. Cheraghi, Business process reengineer-ing: A consolidated methodology, Proceedings of the 4th Annual InternationalConference on Industrial Engineering Theory, Applications, and Practice,1999, US Department of the Interior-Enterprise Architecture (2006).
28. A. V. Looy, M. D. Backer, and G. Poels, Enterprise Information Systems 8, 188(2014).
29. S. Bagchi, X. Bai, and J. Kalagnanam, Data Quality Management Using Busi-ness Process Modeling, Google Patents (2006).
30. W. M. V. D. Aalst, A. H. T. Hofstede, and M. Weske, Business processmanagement: A survey, Business Process Management, Springer (2003),pp. 1–12.
31. W. Bandara, D. R. Chand, A. M. Chircu, S. Hintringer, D. Karagiannis, andJ. C. Recker, Communications of the Association for Information Systems27, 743 (2010).
32. R. K. Ko, S. S. Lee, and E. W. Lee, Business Process Management Journal15, 744 (2009).
33. K. M. Hüner, M. Ofner, and B. Otto, Towards a maturity model for corpo-rate data quality management, Proceedings of the 2009 ACM Symposium onApplied Computing (2009), pp. 231–238.
34. A. Even and G. Shankaranarayanan, Journal of Data and Information QualityJDIQ 1, 15 (2009).
Received: 8 September 2014. Accepted: 12 October 2014.
Dadang Syarif Sihabudin Sahid1�2�∗, Lukito Edi Nugroho1, Ridi Ferdiana1, and Paulus Insap Santosa1
1Department of Electrical Engineering and Information Technology, Universitas Gadjah Mada, Yogyakarta, 55281, Indonesia2Department of Computer, Politeknik Caltex Riau, Pekanbaru, 28265, Indonesia
Currently, the emerging of ubiquitous computing has provided a significant opportunity for using a mobile tech-nology as a learning medium. This situation has led many researchers to investigate the potential of adaptivelearning through mobile learning. However, there are limited researches present a model of adaptive mobilelearning for learners who are always traveling. This paper proposed a context-aware mobile learning modelfor learners who traveling frequently by considering several challenges in traveling situation. This model aimsto provide alternative learning by providing learning resources as well as the location presenting the nearestlearning resources tailored to conditions and situations context of learners. Learning style (video, audio, text,animation), learning preference (case-study, conceptual, simulation) and preferred time are selected as staticcontext parameter and used as an adaptive filter to presenting learning materials through mobile. On the otherhand, location and environment level (connectivity, noise, illumination) are chosen as the dynamic context toprovide the position of the nearest learning resources such as libraries and classmates for collaboration. As theresult of this model provided a wireframe design of context aware mobile learning for the traveler, by providingand recommending the learning materials, learning approach, as well as a nearest source of learning accord-ing to the conditions and situations of learners. In conclusion, this model can provide alternative guidelines fordevelopers who interested implementing context aware mobile learning, especially for frequent traveling learnerto keep earning and following the learning process wherever located.
Keywords: Mobile Learning, Context Aware, Context of Learning, Context of Mobile, Traveling Learner.
1. INTRODUCTIONEnhancing and delivering learning paradigm has been shifted
from traditional all-size type of learner to adaptive and personal-
ized learning. The last paradigm tailors individual learner’s needs
that are fitted with situation, preferences and context of learners.1
These needs include flexibility and mobility for the learner.
In addition, increasing capability the devices and data commu-
nication providers offered opportunities to access the internet
easily. This situation led researchers continuously investigating
and developing the learning activities via mobile device.
Mobile learning enacts feasible correlation between the mobile
device technology and the process of education. As mobile
devices and other portable transmission devices extent persis-
tently, the real worth of mobile learning can be examined in the
implementation of educational.2
Currently, using mobile in the learning process (mobile learn-
ing) gives advantages for the learners, especially for who has
mobile activities. The learners can adapt and personalize the
learning activities due to traveling schedule, place and per-
sonal situation. Many researchers have been experimented on
adaptation and personalization in mobile learning. The research
focused on how the context aware technology supports adaptive
∗Author to whom correspondence should be addressed.
and personalized mobile learning in ubiquitous and pervasive
computing.3�4 Context aware mobile learning systems capable
for providing the most suitable for presenting recommendations
and solutions to learners based on their learning styles and
preferences.
Adaptive and personalized mobile learning has been impor-
tant and desired complex research as well as issues in practices.
It should be considered that without proper functional model
design, integrating contextual information, adaptation and per-
sonalization process as well as presenting adaptation results as
appropriate recommendations for the learners can be compli-
cated and hard to be carried out. Moreover, the complexities
of the needs and preferences, situation, learning environment,
processes, included implementation methods increases drastically
when taking into account the support of mobile learners with
various of mobile devices accessing adaptive and personalized
recommendations. However, existing researches on context aware
mobile learning for traveler are still limited due to some issues
and challenges such as in designing, processing, communicating,
presenting previous e-learners to mobile learners, and traveling
situation contexts.5
This paper presented a context aware mobile learning model
for the traveler by describing the design of contextual learner,
adaptation process and presenting adaptation results. The model
rithm, and similarity algorithm, to proceed particular context of
learning or context of mobile.
4.3. Adaptation of Context Aware Mobile Learning
There are two primary classifications of adaptation in context
aware mobile learning: (1) adaptation related to resources of
learning; (2) adaptation related to activities of learning.25 Each of
classifications is divided into sub-category or kind of adaptations.
The first classification consisted of:
• Selecting, it is carried out by filtering relevant materials of
learning and provided to the learners based on context of learning
and context of mobile.
• Presenting, it is how to show the learning material through
mobile devices.
• Navigating and sequencing, it is rearranging or reordering
resources to make the individual learning pathways.
The second classification of adaptation context aware mobile
learning consisted of:
• General adaptation, it is generating activities of per-
sonal learning automatically, according to several require-
ments captured from context of learning and context of
mobile.
• Feedback and support, it is an adaptation by providing individ-
ual advises and suggestions related to learning time and learning
activities.
Table II. Dimension elements of context.24
Dimensions Contexts
Context of learningDesign of learning Objectives of learning, pedagogical approaches,
activities of learning, roles of participation,tools, and resources of learning.
Profile of learner Expertise (skills, attitudes, knowledge), role,individual condition and characteristics(learning needs, learning styles, learninginterests, physical condition or other inabilities).
Context of mobileLearner Temporal and situational condition (preferences,
mood, needs, interests).People Role, relationship, contributions and constraints.Place Location and position, zones, room of interactive,
setting and culture of learning.Artifact Technological and non-technological aspect.Time Duration of task, schedule of task, availability,
and action happens.Environment Network quality, illumination, noise level, weather
conditions.
3072
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3070–3074, 2015
• Navigation to locations, it is and adaptation by providing
awareness of place and position related to appropriate activities
of learning.
• Communication and interaction, it is an adaptation by facili-
tating the learners while executing activities of learning.
5. THE PROPOSED MODEL OF CONTEXT
AWARE MOBILE LEARNING FOR
TRAVELERPrevious concept, related works, tools and design model for con-
text aware adaptive and personalized mobile learning has been
given suitable and proper foundation to be implemented in con-
text aware mobile learning for traveling learner. The concept
of mobile learning gives basic knowledge and characteristics of
mobile learning. Related works in mobile learning application
shows the state of the art about researches in adaptive and per-
sonalized context aware mobile learning, whereas recommender
mobile applications in tourism gives inspirations, how the con-
text aware mobile learning implemented for travelers especially
how to deal with the constraints of traveling situation context
such as limited connectivity, noise and illumination level. Design
adaptation of context aware mobile learning gives guidance and
become tools to develop the proposed model. Figure 3 shows the
components and elements for the proposed context aware mobile
learning model for traveler.
In the learner contextual information, the proposed model
using learning styles, learning preferences, and time preference
as learning context that is input by the learner. Location with
GPS technology sensor is used as mobile context to detect near-
est library and classmates location. As well as the Wi-Fi, noise
and illumination sensor will be used by the system to select
appropriate learning materials. In adaptation engine phase, adap-
tation rules such as condition structure rule based are imple-
mented as the approach. In this phase, also are provided with
and resources database. In the adaptation phase, some adapta-
tion types are conducted. Selection adaptation type is used for
Fig. 3. Components and elements of the proposed model.
Fig. 4. Wireframe design of the proposed model.
presenting learning material recommendations based on learning
style and learning preferences by dealing with the connectivity
level (poor, good, strength), noise level, and illumination level.
On the other hand, navigation to location, communication and
interaction adaptation type are used for presenting library and
nearest classmates location.
3073
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3070–3074, 2015
In order to give more description about the proposed model,
this paper also provides simple wire frame design as can be
shown in Figure 4. The wire frame design is a low fidelity pro-
totype design to figure out the model if will be implemented.
6. CONCLUSIONThe proposed model of context aware mobile learning for the
traveler has been presented in this paper. The model has been
developed based on adaptive design as a tool of adaptive and
personalized context aware mobile learning. The tool has given
suitable knowledge for building a context aware mobile learning.
Traveling learner is chosen as a model implementation subject
due to limitation of researches and works in this area due to
some issues and challenges related to traveling situation contexts.
All processes in building the proposed model has been carried
out as a foundation for implementation. The proposed model is
expected not only improving the previous context aware mobile
learning model, but also contributing to learners especially who
have traveling activities frequently. In the future works, the model
can be implemented in a real situation by considering more com-
plex of the contexts, adaptation engines, and type of adaptations.
References and Notes1. S. Gómez, P. Zervas, D. G. Sampson, and R. Fabregat, Delivering adaptive
and context-aware educational scenarios via mobile devices (2012).2. Y. Jiugen, X. Ruonan, and W. Jianmin, Applying research of mobile learning
mode in teaching 417 (2010).3. K. Chin and Y. Chen, The 2nd International Conference on Integrated Informa-
tion a Mobile Learning Support System for Ubiquitous Learning Environments,Procedia—Soc. Behav. Sci. (2013), Vol. 73, pp. 14–21.
4. A. Syvänen, R. Beale, M. Sharples, M. Ahonen, and P. Lonsdale, Support-ing pervasive learning environments: Adaptability and context awareness inmobile learning 5 (2005).
5. Y. Chia and F. S. Tsai, Context-aware mobile learning with a semantic service-oriented infrastructure 896 (2011).
6. S. Karadeniz, Flexible design for the future of distance learning, Procedia—Soc., Behav., Sci. (2009), Vol. 1, pp. 358–363.
7. I. Conference and D. Learning, M-learning: An innovative advancement of ICTin education 148 (2010).
8. S. Chin, Mobile technology and gamification: The future is now! 138(2014).
9. Y. Jin, Research of one mobile learning system 162 (2009).10. S. Behera, International Journal on New Trends in Education and Their Impli-
cations 65 (2013).11. S. Wu, A. Chang, M. Chang, T. Liu, and J. Heh, Identifying personalized
context-aware knowledge structure for individual user in ubiquitous learningenvironment 95 (2008).
12. R. M. Carro, Supporting the development of mobile adaptive learning envi-ronments: A case study 2, 23 (2009).
13. J. Y. Yau and M. Joy, A context-aware personalized m-learning applicationbased on m-learning preferences.
14. S.-L. Wang and C.-Y. Wu, Expert Syst. Appl. 38, 10831 (2011).15. F. Ako-nai, Q. Tan, and F. C. Pivot, The 5R adaptive learning content gener-
ation platform for mobile learning (2012).16. P. Zervas and D. G. Sampson, Context-aware adaptive and personalized
mobile learning delivery supported by UoLmP 47 (2014).17. C.-C. Yu and H. Chang, Towards Context-aware recommendation for per-
sonalized mobile travel planning, Context-Aware Systems and ApplicationsSE-12, edited by P. Vinh, N. Hung, N. Tung, and J. Suzuki, Springer, Berlin,Heidelberg (2013), Vol. 109, pp. 121–130.
18. D. Gavalas, C. Konstantopoulos, and K. Mastakas, J. Netw. Comput. Appl.39, 319 (2014).
19. J. Borràs, A. Moreno, and A. Valls, Expert systems with applications intelligenttourism recommender systems: A survey 41, 7370 (2014).
20. M. F. Fudzee, A classification for content adaptation system 426 (2008).21. J.-M. Su, S.-S. Tseng, H.-Y. Lin, and C.-H. Chen, User Model. User-Adapt.
Interact. 21, 5 (2011).22. P. Zervas, D. Sampson, S. Gómez, and R. Fabregat, Designing tools
for context-aware mobile educational content adaptation, Ubiquitous andMobile Learning in the Digital Age SE-3, edited by D. G. Sampson,P. Isaias, D. Ifenthaler, and J. M. Spector, Springer, New York (2013),pp. 37–50.
23. M. M. Das, Static context model for context aware E-learning 2, 2337 (2010).24. P. Zervas, S. Eduardo, G. Ardila, R. Fabregat, and D. G. Sampson, Tools for
context-aware learning design and mobile delivery 3 (2011).25. D. Sampson and P. Zervas, Context-aware adaptive and personalized mobile
learning systems, Ubiquitous and Mobile Learning in the Digital Age SE-1,edited by D. G. Sampson, P. Isaias, D. Ifenthaler, and J. M. Spector, Springer,New York (2013), pp. 3–17.
Received: 25 September 2014. Accepted: 29 October 2014.
Chairul Saleh1�∗, Raden Achmad Chairdino Leuveano2, Mohd Nizam Ab Rahman2,Baba Md Deros2, and Nur Rachman Dzakiyullah3
1Department of Industrial Engineering, Faculty of Industrial Technology, Universitas Islam Indonesia,Yogyakarta, 55584, Indonesia
2Department of Mechanical and Material Engineering, Faculty of Engineering and Build Environment,Univeriti Kebangsaan Malaysia, Malaysia
3Department of Informatics Engineering, Faculty of Engineering, Janabadra University, Yogyakarta, Indonesia
In this paper, the back-propagation artificial neural networks (ANN) model is presented to predict expenditureof carbon (CO2) emission. The model was built based on the input variables that affect to expenditure of CO2
include the amount of bagasse, wood and marine fuel oil used in boiler machine. The objective of this paperis to monitor the CO2 emission based on the fuel used for operating the boiler machine. The data used fortesting the models were obtained from Sugar Industry. It splits up into 90% of training data and 10% of testingdata. The model experiment was conducted using trial and error approach to find the optimal parameters ofANN model. The result shows that the architecture of ANN model have optimal parameter on training cycle 50,learning rate 0.1, momentum 0.1, and 19 hidden nodes. The validity of the trained ANN is evaluated by usingRoot Mean Square Error (RMSE) with error value as 0.055. It indicates that the smallest error provides moreaccurate results on prediction and even can contribute to the industrial practice, especially helping the executivemanager to make an effective decision for business operation by considering the expenditure of CO2 monitoring.
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3080–3084, 2015
residue), firewood and marine fuel oil (MFO). Monitoring these
fuels is very important because they have different effects on the
emission of CO2.
To monitor machine emissions, most studies use regression,
discriminant analysis and artificial neural networks (ANN) in
empirical modelling to control the system.6–8 Of these, the ANN
model has proved the most attractive to researchers for pre-
dicting the behaviour of a system in certain cases and neural
networks have increasing been used in recent years over conven-
tional statistical techniques such as regression and discriminant
analysis.9 The model includes machine performance testing, cut-
ting mechanics, signal processing, data decomposition and image
processing.10 The ANN model is a statistical learning algorithm,
the design of which was inspired by the properties of biologi-
cal neural systems, used to search and produce new knowledge
to estimate functions based on a large number of inputs.11 The
ANN model can also be defined as a mathematical model for
predicting new problems.12
As harmful emissions have increased, previous research has
also attempted to embed the use of ANNs in environmental appli-
cations. Baareh13 applied the ANN model to estimate CO2 emis-
sions for consumption from four fuel inputs, including global oil,
natural gas (NG), coal and primary energy (PE). Fontes et al.14
employed ANN model to classify ozone episodes, which have a
negative impact on the environment. This aimed to reduce ineffi-
ciency leading to ozone precursor emissions. To prevent fouling
problems in machines resulting in higher CO2 emissions, Romeo
and Gareta15 and Rusinowski and Stanek16 used the ANN method
to monitor boiler performance. As the ANN model is a powerful
tool for handling such types of modelling processes, this paper
proposes an ANN model to predict CO2 emissions in boilers in
the sugar industry. The objective is to monitor CO2 emissions
based on fuel combustion used to operate the boilers in sugar
production. By monitoring the machines, the sugar industry can
efficiently manage fuel combustion.
The remainder of this paper is organized as follows: Section 2
introduces the materials and methods in the design of the ANN
model; Section 3 presents the result of the experiment; finally,
Section 4 summarizes the salient points and concludes the paper.
2. MATERIALS AND METHODSTo design the ANN model for predicting the CO2 emissions from
boiler use, a number of steps were defined as shown in Figure 1.
The first step to in the design of the ANN model is data col-
lection. The primary data were collected from the sugar indus-
try in Yogyakarta. This research analyses boiler emissions based
on the fuel used. The three main fuels that affect CO2 emis-
sions are found to include bagasse, firewood and MFO. These
fuels are thus categorized as input variables. The output vari-
able (CO2 emission) is computed by multiplying activity data
(e.g., fuel consumed) by the emission factor for that activity,
in accordance with the guidelines for computation of emissions
provided by the IPCC.17 The relationship between the input and
output variables is defined as:
Xi =
∣∣∣∣∣∣∣Bagasse�X1
Firewood�X2
MFO�X3
∣∣∣∣∣∣∣ � Y = CO2 (1)
Dataset cleansing, transformation-normalization
Cross-validation
Training Set Test Set
Data collection
• Bagasse• Firewood• MFO• Carbon emission
12
10....................................
Total Data
k-fold cross-validation
Model Search
Modelevaluation
Fig. 1. Design overview of the ANN model.
To avoid any noisy data, missing data, incorrect, improp-
erly formatted, or duplicated in datasets, then data cleansing is
employed.22 The aim of data cleansing is to have a better repre-
sentative datasets for developing reliable neural network model
and improve the accuracy of prediction. The process of data
cleansing in this study was automatically performed by Rapid
Miner software. However, as shown in Figure 2, the cleansing
process shows that dataset has zero noisy and missing value. The
next step is to transform and normalize the dataset in order to
have inputs with 0 means and a standard deviation of 1.
Then, to validate the model, k-fold cross-validation is used.
This technique divides the dataset into a training set and a test
set. The training set is used to calculate the gradient and update
the network weights and biases. In this case, validation was com-
puted during the training process to obtain minimum error in
prediction. The test dataset is used to assess the performance of
Fig. 2. The process of data cleansing using rapid miner.
3081
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3080–3084, 2015
a fully-trained ANN model with data outside the training set.
To find the optimal architecture of the ANN model in the training
process, the parameters—training cycle, learning rate, momen-
tum and hidden node—must be optimal. The simplified architec-
ture of the ANN model comprising three layers (input variables,
hidden nodes and output variable) is shown in Figure 3.
A back-propagation (BP) learning algorithm is then used in the
training process to obtain the optimal parameters. BP training is
categorized as a gradient descent algorithm.7 With this technique,
the total error can be reduced by changing the weight along its
gradient. However, to analyse the effect of network parameter,
a trial and error approach is used in the design of the BP neural
network prediction by varying the network structure based on the
pre-processing of the data, the number of input nodes and the
activation function. To evaluate the model error, the root mean
square error (RMSE) is expressed by:
RMSE = 1
2
[∑j
∑k
�Yjk −Ojk�]1/2
(2)
where Y is the predicted value and O denotes the actual value
vectors over pattern k.
3. RESULTSRapid Miner 5.2.003 was to carry out the experiment. All the test
problems were undertaken using a computer with the following
specifications: Intel (R) Core (TM) i5-2450M CPU @ 2.50 GHz
(4 CPUs), 8 GB of RAM and Windows 8.1, 64-bit (6.3, Build
9600) operating system. As mentioned previously, the data for the
experiment were obtained from the sugar industry in Yogyakarta,
Indonesia. There are 124 datasets for the period 2009–2013 on
the use of fuels, including bagasse (tonnes), firewood (tonnes)
Fig. 3. Simplified architecture of the ANN model, adopted from Erchan andAtici.20
Table I. Optimal parameters of the back-propagation ANN model.
Parameters Value
Training cycle 50Learning rate 0�1Momentum 0�1Hidden nodes 19RMSE 0�055
and MFO (litres), for boiler operation. From the data on fuel
consumed, the next step is to calculate the CO2 emissions (tonnes
CO2) using the IPCC17 (2006) guidelines. Our objective is to
monitor CO2 emissions from boilers, seek an accurate prediction
that has the lowest error. In other words, an accurate prediction
can provide information regarding the fuel consumption that has
the lowest emissions.
In this experiment, 10-fold cross-validation was used to split
the data into a training set and a test set. The data were divided
into 10 sets of size n/10 or equal parts: 90% of the data were
used for the training set and 10% for the testing set. Error evalu-
ation was then performed on 90/10 splits, repeated for all 10 pos-
sible splits. For each repetition, the training fold was normalized.
The mean values and standard deviations were taken over k dif-
ferent partitions. The k-fold cross-validation technique was used
as it can provide a lower variance, meaning that minimum error
can be achieved.
As noted above, the ANN model consists of three layers, for
which the parameters include the training cycle, the learning rate,
momentum and hidden nodes. Using the BP algorithm with a trial
and error approach for the training process, the optimal architec-
ture of the ANN model is as shown in Table I.
The objective of the training process is to minimize the error
of prediction. This is taken as a rule to choose the optimal param-
eters of the ANN model. As can be seen from Table I, the perfor-
mance of the ANN model has an error value (RMSE) of 0.055.
The closer the error value is to zero, the higher the accuracy of
prediction. When the error value achieved its minimum, the train-
ing cycle was terminated at 50 cycles. The learning rate of the
ANN model was 0.1. The value of the learning rate represents
0.000
0.200
0.400
0.600
0.800
1.000
1.200
1 9 17 25 33 41 49 57 65 73 81 89 97 105
113
121
Err
or V
alue
Data
Actual
Prediction
Fig. 4. Comparison of actual values versus predicted values.
3082
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3080–3084, 2015
Table II. Test of normality.
Shapiro–Wilk
Statistic df Sig.
0.977 124 0.032
the speed at which the system learns (converges). The momen-
tum rate has the same value as the learning rate, which is 0.1.
The momentum rate is used to avoid local minima and significant
changes in weights so that the global minimum can be achieved.
The optimal hidden nodes in the network numbered 19. Deter-
mining the optimal number of hidden nodes can avoid overfitting
and underfitting of the model. Overfitting occurs when the learn-
ing algorithm captures noise from the data, whereas underfitting
occurs when the learning algorithm cannot correctly detect the
underlying trend of the data, so that the model shows low bias but
high variance. Based on the results of the experiment, Figure 4
shows the actual and predicted values for both training and test
cases.
In the last step, the model output of the neural network is
analysed. This analysis aims to determine the level of confidence
in the trained model by seeing whether the predicted values are
normally distributed.18 If the predicted value is normally dis-
tributed, the model prediction has higher accuracy and precision.
Several statistical tests can be employed to check the normal-
ity of distribution, including the Shapiro–Wilk test, the Lilliefors
test, D’Agostino–Pearson’s L2 test, the Jarque–Bera test, and the
Anderson–Darling test. However, the most applicable statistical
method that fits all types of distribution and sample size is the
Shapiro–Wilk test19 and thus it was used in this study to test nor-
mality. In this paper, the significance level (�) was set at 0.05; if
p < 0�05, the data are not normally distributed. The result of the
Shapiro–Wilk test is shown in Table II.
As can be seen from Table II, the significance is 0.032. Thus,
the predicted values are not normally distributed. However, if
the statistical test does not show normal distribution, histogram
analysis can be used, as shown in Figure 5.
Basically, statistical data analysis is used to recognize the sta-
tistical probability distribution of the data. As the result, statisti-
cal inference and information based on data can be obtained and
Fig. 5. Histogram of normality distribution.
help to make the right decision.21 From Figure 5, it transpires
that despite the significance value found above, the prediction
results approximate normal distribution. Although, the prediction
model has higher accuracy that shown by lower RMSE value,
however, the histogram is centred over the true value. The result
shows the prediction model has poor repeatability and poor pre-
cision. However, the output of the ANN model is still proven to
solve non-linear data and provide higher prediction accuracy.
4. CONCLUSION AND RECOMMENDATIONThe back-propagation ANN model was applied in this paper to
predict CO2 emissions from boiler operations. The prediction
model is used to monitor fuel combustion from bagasse, fire-
wood and MFO, which affect the amount of CO2 emitted. The
ANN model was designed with three layers (input variable, hid-
den nodes and output variable). To obtain better prediction with
a lower error (RMSE) value, the trial and error approach was
applied. The minimum error value can also be used to optimize
the parameters of the ANN model. The results obtained show
that the RMSE value was 0.055 with optimal parameters for the
ANN model of 50 for the training cycle, 0.1 for the learning
rate, 0.1 for the momentum rate and 19 for the number of hidden
notes.
Greater accuracy in prediction can provide accurate informa-
tion regarding CO2 emissions. It means that when creating the
model prediction, the main goal is to achieve the lower RMSE
value. It can be conclude that lower RMSE value, greater accu-
racy of prediction can be obtained. As the result of this predic-
tion, this study can help manager to monitor boiler machines and
then develop policies or take decisions regarding production that
reduce the negative impact on the environment. By this reason-
ing, this study can contribute to practice in monitoring machine
emissions using ANN model prediction. In further research, this
ANN model could be integrated with optimization techniques,
such as genetic algorithms, particle swarm optimization and ant
colony optimization to improve the accuracy of prediction.
Acknowledgment: The authors would like to thank the
anonymous reviewers for their valuable comments. This research
was supported by This research is supported by Directorate of
Research and Community Service and Board Academic Devel-
opment, Universitas Islam Indonesia.
References and Notes1. IPCC, Climate change 2014: Mitigation of climate change [Online] Available
from: http://report.mitigation2014.org/report/ipcc_wg3_ar5_full.pdf [Accessedon 16th August 2014] (2014).
2. IPCC, Climate change 2007: The physical science basis [Online] Avail-able from: http://www.ipcc.ch/pdf/assessment-report/ar4/wg1/ar4-g1-spm.pdf[Accessed on 16th August 2014] (2007).
3. WRI, WRI’s climate data explorer: GHG emissions-energy-sub-sector [Online]Available from: http://cait2.wri.org/ [Accessed on 16th August 2014] (2011).
4. Industrial Department, Roadmap industri gula [Online] Available from:http://agro.kemenperin.go.id/e-klaster/file/roadmap/ [Accessed on 16th August2014] (2009).
5. H. T. Pao and C. M. Tsai, Energy 36, 2450 (2011).6. Y. Çay, A. Çiçek, F. Kara, and S. Sagiroglu, Applied Thermal Engineering
37, 217 (2012).7. S. A. Kalogirou, Progress in Energy and Combustion Science 29, 515 (2003).8. A. Maijidan and M. H. Zaidi, International Journal of Fatigue 29, 489
(2007).9. M. Paliwal and U. A. Kumar, Expert Systems with Applications 36, 2 (2009).
3083
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3080–3084, 2015
10. A. Negarestani, S. Setayeshi, M. Ghannadi-Maragheh, and B. Akashe,Applied Radiation and Isotopes 58, 269 (2003).
11. H. Yoon, C. S. Park, J. S. Kim, and J. B. Baek, Expert Systems with Applica-tions 40, 231 (2013).
12. G. Zhang, B. E. Patuwo, and M. Y. Hu, International Journal of Forecasting14, 35 (1998).
13. A. K. Baareh, Journal of Software Engineering and Applications 6, 338 (2013).14. T. Fontes, L. M. Silva, M. P. Silva, N. Barros, and A. C. Carvalho, Science of
the Total Environment 488, 197 (2013).15. L. M. Romeo and R. Gareta, Applied Thermal Engineering 26, 1530 (2006).16. H. Rusinowski and W. Stanek, Energy Conversion and Management 48, 2802
(2007).
17. IPCC, IPCC guidelines for national greenhouse gas inventories. Energy (2)[Online] Available from: http://www.ipcc-nggip.iges.or.jp/public/2006gl/vol2.html [Accessed on 16th August 2014] (2006).
18. D. K. Williams, Jr., A. L. Kovach, D. C. Muddiman, and K. W. Hanck, AmericanSociety for Mass Spectrometry 20, 1301 (2009).
19. N. M. Razali and Y. B. Wah, Journal of Statistical Modelling and Analysis 2, 21(2011).
20. S. Ercan and U. Atici, Neural Computation and Application 22, 1039(2013).
21. C.-T. Su and C.-J. Chou, Quality Engineering 18, 293 (2006).22. C. Vercellis, Business Intelligence: Data Mining and Optimization for Decision
Identification E-Learning Readiness in the Faculty
of Agricultural Technology Jambi University
Kurniabudi∗, SetiawanAssegaff, and Sharipuddin
STIKOM Dinamika Bangsa Jambi Jl. Jenderal Sudirman, Thehook, Jambi, Indonesia
Identification of E-learning readiness needs to be done so that the implementation of e-learning can workwell without spending money, time and effort. This study aims to identify the level of readiness of e-learningat the Faculty of Agricultural Technology Jambi University (abbreviated as FTP UNJA). Collecting data usingquestionnaires. Identify the level of e-learning readiness ELR model. The study concluded FTP UNJA be in aposition ready for the implementation of e-learning with some improvement in the perception of lecturers andmanagement policies.
(score 3.852) and management support (a score of 3,500). On
average each factor has a value score of >3.4, it indicates FTP
UNJA be in a position ready for the implementation of e-learning
with some improvement.
From cultural factors, especially the lectures perception
towards e-learning, FTP UNJA are at the ready position with a
score of 3,852. This means that each lecture has a positive per-
ception of the use of e-learning. Which needs to be addressed
is the perception that e-learning was difficult to use, particularly
the use of e-learning tools (a score 3.250).
Although based management support has a mean score of
3,500, which means being in the ready position, but one of the
aspects that are part of management support is the support of
e-learning policies and regulations on intellectual property rights
still be in a position not yet ready.
Acknowledgments: This research is supported and funded
by STIKOM Dinamika Bangsa through an internal faculty
research grants in 2014-2015.
References and Notes1. Al-Radhi, A. Al-Din, and J. Kadhem, Bulletin of the American Society for Infor-
mation Science and Technology (Online) 34, 3 (2008).2. A. A. Al-Furaydi, English Language Teaching 6 (2013).3. A. Saekow and D. Samson, International Journal of e-Education, e-Business,
e-Management and e-Learning 1 (2011).4. C. H. Aydin and D. Tasci, Educational Technology and Society 8, 244 (2005).5. S. Borotis and A. Poulymenakou, E-learning readiness components: Key
issues to consider before adopting e-learning interventions, edited by J. Nalland R. Robson, Proceedings of World Conference on E-Learning in Cor-porate, Government, Healthcare, and Higher Education 2004, Chesapeake,VA, AACE Retrieved August, 2014 from http://www.editlib.org/p/11555 (2004).pp. 1622–1629.
6. S. Chapnick, Are you ready for e-learning, Retrieved August 2014, fromhttp://blog.uny.ac.id/nurhadi/files/2010/08/are_you_ready_e for_elearning.pdf.
7. R. C. Clark and R. E. Mayer, e-Learning and the Science of Instruction:Proven Guidelines for Customers and Designers of Multimedia Learning,Third edn., San Francisco, Pfeiffer, CA (2008).
8. T. Eslaminejad, M. Masood, and N. A. Ngah, Assessment of Instructors’Readiness for Implementing e-Learning in Continuing Medical Education inIran, University of Medical Sciences, Kerman, Iran (2010).
9. H. M. Azimi, Journal of Novel Applied Sciences 769 (2013).10. D. Haney, Performance Improvement 41, 8 (2002).11. K. Kaur, and Z. W. Abas (2004), An assessment of e-learning readiness at
Open University Malaysia, International Conference on Computers in Educa-tion (2004).
12. Oketch and H. Achieng, E-Learning Readiness Assessment Model In Kenyas’Higher Education Institutions: A Case Study Of University Of Nairobi (2013).
13. P. Sarantos, Presumptions and actions affecting an e-learning adoption by theeducational system. Implementation using virtual private networks (2005).
14. M. J. Rosenberg, The E-Learning Readiness Survey, Retrieved August2014, from http://www.books.mcgraw-hill.com/training/elearning/elearning_survey.pdf (2000).
15. P. Saunbang and P. Petocz, International Journal of E-Learning 5, 415 (2006).16. K. B. Trombley and D. Lee, Journal of Educational Media 27, 137 (2002).17. R. Vilkonis, T. Bakanoviene, and S. Turskiene, Readiness of Adults to Learn
Using E-Learning, M-Learning and T-Learning Technologies, Informatic inEducation Vilnius University (2013), Vol. 12, pp. 181–190.
Received: 30 October 2014. Accepted: 20 December 2014.
Department of Civil Engineering, Janabadra University, Yogyakarta, Indonesia
The study on castellated beam with hexagon shaped web opening has been carried out, so the crack initiationat the opening-corners could easily be determined. In relation to that, the development of structural optimiza-tion with design variable is considerably needed, i.e., layout and opening size of non-angular side at I-sectionsteel beam with circular or cellular web-opening shape. Genetic Algorithm (GA) has been identified to solvedesign optimization problem. The main target is determine optimum shape, size and layout opening with cellularI-section steel beam, so the basis criterion of shape and size planning of castellated steel beam opening canbe obtained. From the data analysis of optimization using numerical computation, it can be obtained that themaximum loading capacity with even number of openings loaded in centric achieves 10%–15% compared to theodd number of openings. The castellated cellular beam of I-Section beam can increase its capacity to 3.5 timesthan the initial beam. Parametric study of the opening dimension and layout analysis with the opening heightfrom the centric with maximum cutting 0.65 hw is in the range of 0.25–0.60 from the width of openings. It canalso be known that the optimum shapes are circle and vertical ellips. It is hoped that results of this study mayfuther develop the current tradition of designing castellated steel beam.
The purpose of mutation is to avoid local minima by prevent-
ing the population chromosomes, which may slow down or even
completely stop to process of evolution prematurely. This is
also avoid only taking the fittest chromosomes is generating the
next population, but rather adopting a weighted random selec-
tion toward that those are fitter. There is, however, one parameter
which is of vital importance in mutation, i.e., the mutation prob-
ability Pm. It control the number of mutated genes that needs
evaluation. Too small mutation probability would overlook useful
possible genes. The resulting mutated genes need to be investi-
gate for their acceptability, i.e., if they are still in the solution
domain or otherwise. A refinement process may applied to genes
which are unacceptable. The example of mutation is shown in
Figure 6.
4.7. Termination
In this step, termination of GA process may be difficult to iden-
tify convergence criteria. Based on Pasandideh et al.18�19 the cri-
teria to stop the generation is either by stopping after a fixed
number of generations or significant improvement in the solu-
tion. It is produce by comparing the average fitness value of the
current generation with that of the preceding one. This paper
conducts a fixed 200 generations to search the solution.
4.8. Population Generation
Generation of the population is resulted by the initial chromo-
some that acquires new chromosome called offspring through the
process of crossover and mutation. The descending process is
then conducted to the initial and new chromosome, followed by
determining the most fitness chromosome which is having the
most optimal value. Figure 7 show result the optimal value by
presenting the average, best and worst fitness value.
Table I, to obtain maximum fitness function takes the weight
percentage of the value of C1 (parameter area) and C2 (stress
parameters). Analysis of the simulations carried out on the open-
ing are numbered and odd that give effect to the value of C1
and C2. The objects of experiment is steel beams Profiles I elastic
area with specimens of 150×75×5×7 mm.
Table II, recapitulates the resulting carrying capacity than the
odd number of openings. More specially,
a. The increase of elastic strength is around 10%, if the even
number of openings is larger than the odd one
b. The increase in strength is around 26%, if the even number
of opening is less than the odd one.
5. CONCLUSIONS AND RECOMMENDATIONThe research reveals that
a. Optimized cellular beams with even number of web openings
subjected to load patterns considered in this research has greater
elastic carrying capacity than those odd number of web opening
b. Optimization result show that the elastic carrying capacity of
cellular beam of equal length increase sharply if the even number
of openings is smaller than that one of the odd one
c. Optimized cellular beams with even number of web openings
subjected to load patterns considered in this research has greater
elastic carrying capacity than those odd number of web opening
d. Optimized cellular beams with even number of web openings
subjected to load patterns considered in this research has greater
elastic carrying capacity than those odd number of web opening
e. Optimization result show that the elastic carrying capacity of
cellular beam of equal length increase sharply if the even number
of openings is smaller than that one of the odd one
f. Cellular opening produce by optimization have forms close to
circles for short beams and produced upright ellipses for longer
span ones
3179
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3176–3180, 2015
g. Optimum elastic carrying capacity of cellular beams with
one point concentrated load is controlled by allowable stresses
for short span and it is controlled by deflection for long span
ones
h. Elliptical or circular cellular web opening with smooth sides
eliminate stress concentration taking place in the corners of
hexagonal openings
i. The castellated cellular beam of I-Section beam can increase
its capacity to 3.5 times than the initial beam.
References and Notes1. Anon, Partial Differential Toolbox for Use with MATLAB-User’s Guide, The
Mathwork, Inc., 3 Apple Hill Drive, Natick, MA (2006).2. A. Chipperffied, P. Fleming, H. Pohlheim, and C. Fonseca, Genetic algorithm
toolbox for use with MATLAB, User guide, Department of Automaic Controland System Engineering, University of Sheffield (2004).
3. Kerdal and D. A. Nethercot, J. Constr. Steel Res. 4, 295 (1984).4. W. Zaarour and R. Redwood, J. Struct. Eng. 122, 860 (1996).5. R. Redwood and S. Demirdjian, J. Struct. Eng. 124, 1202 (1998).6. R. Delesquez, Metallique 3, 26 (1968).7. A. A. Aglan and R. G. Redwood, Web buckling in castellated beams, ICE
Proceedings (1974), Vol. 57, pp. 307–320.
8. R. M. Lawson, Design of FABSEC Cellular Beams in Non-composite andComposite Applications for Both Normal Temperature and Fire EngineeringConditions to SCI AD 269, Fabsec Limited. (2004).
9. E. Ellobody, Thin-Walled Struct. 52, 66 (2012).10. P. Wang, X. Wang, and M. Liu, Thin-Walled Struct. 85, 441 (2014).11. S. Durif, A. Bouchaïr, and O. Vassart, Eng. Struct. 59, 587 (2014).12. S. Chen, T. Limazie, and J. Tan, J. Constr. Steel Res. 106, 329 (2015).13. P. Wang, X. Wang, M. Liu, and L. Zhang, Thin-Walled Struct. (2015).14. D. Sonck and J. Belis, J. Constr. Steel Res. 105, 119 (2015).15. M. Gen and R. Cheng, Genetic Algorithm and Engineering Optimization, John
Wily Sons, New York (2000).16. C. Saleh, V. Avianti, and A. Hasan, Optimization of fuzzy membership func-
tion using genetic algorithm to minimize the mean square error of credit sta-tus prediction, The 11th Asia Pacific Industrial Engineering and ManagementSystems Conference The 14th Asia Pacific Regional Meeting of InternationalFoundation for Production Research (2010).
17. C. Saleh, A. Hassan, B. M. Deros, M. N. A. Rahman, R. A. C. Leuveano, andA. Adiyoga, Parameters optimization of VMI system in a manufacturer andmulti retailer using genetic algorithm, International Conferance on AdvancedManufacturing and Material Engineering 2014 (ICAMME 2014) (2014).
18. S. Silva, A Genetic Programming Toolbox for MATLAB, Evolutionary and Com-plex System Group, University of Coimbra, Portugal (2004).
19. S. H. R. Pasandideh, S. T. A. Niaki, and J. A. Yeganeh, Adv. Eng. Softw.41, 306 (2010).
20. S. H. R. Pasandideh, S. T. A. Niaki, and A. R. Nia, Expert Syst. Appl. 38, 2708(2011).
Received: 27 January 2015. Accepted: 25 March 2015.
Ir. Endah Budi Purnomowati∗, Gaguk Asmungi, Anindito Yusuf Wirawan, and Rudy Yuwono
Department of Electrical Engineering, Faculty of Engineering, University of Brawijaya, Malang 65145, Indonesia
Pathloss is a decrease in the power level that caused by path attenuation of radio waves. Pathloss is dependenton several factors, one of them is the carrier frequency used. At a distance of 0.41 km and a carrier frequencyof 1920 MHz, the resulting loss is 127.41 dB. At the same distance with a carrier frequency of 2110 MHz, theresulting loss of 128.51 dB. So it can be concluded that the higher the carrier frequency used will result on thegreater the loss. To support the demand of the downlink traffic, it would be better if the carrier frequency usedin the downlink carrier frequency is lower than on the uplink.
Keywords: Pathloss, Carrier Frequency, WCDMA.
1. INTRODUCTIONPathloss is a decrease in the power level that caused by path
attenuation of radio waves, such as refraction, diffraction, reflec-
tion, and scattering. Pathloss is very dependent on the distance
of transmitter and receiver antennas and the carrier frequency
used. This study aims to provide analysis of the pathloss calcu-
lation in uplink and downlink traffic using different frequencies
according to the standard that has been designed by the Inter-
national Telecommunications Union (ITU), known as IMT-2000
(International Mobile Telecommunications 2000). It also aims to
give good recommendation on using carrier frequencies for traf-
fic downlink to support customer demands in order to keep it
running optimally. The pathloss calculation in this study using
Walfish-Ikegami model, the distance of transmitter and receiver
used are as far as 0.41 km, 0.43 km, 0.46 km, 0.67 km and
0.82 km.
2. LITERATURE OVERVIEW2.1. 3G (Third Generation)
3G, stands for third generation, is the third generation of mobile
1Department of Computer Science, Faculty of Science at Siracha, Kasetsart University Siracha Campus,Chonburi, 20230, Thailand
2Department of Computer Science, School of Applied Statistics, National Institute of Development Administration,Bangkok, 10240, Thailand
Internet banking has become the norm for many simple banking transactions such as money transfers, goodsand services payments, etc. However, conducting banking transactions via the Internet may be subjected tomany types of attacks including password attacks, malware, phishing, and other unauthorized activities. Manybanks have enhanced their security by using One-Time Password (OTP) as another authentication method inaddition to traditional username and password. The OTP may be sent to the mobile phone number of the accountowner via SMS. Even with the enhanced security measure, the Internet banking is still vulnerable to differenttypes of attacks such as online phishing. We propose, design, and implement two transaction authenticationschemes using mobile OTP and QR Code. Both schemes are resilient to known attacks including, but not limitedto, eavesdropping, replay, message modification, and phishing.
Keywords: OTP, QR Code, Internet Banking, Authentication, Mobile Phone.
1. INTRODUCTIONInternet banking has become the norm for many simple bank-
ing transactions such as money transfers, goods and services
payments, etc. The user simply logs into the account using the
username and password as the credentials through a secure con-
nection. When the user requests a financial transaction such as
money transfer, and goods and services payment, the bank may
require the user to confirm the transaction using another form
of authentication. For example, the bank sends an SMS One-
Time Password (OTP)1 to the user’s previously registered mobile
phone. The user commits the transaction by submitting the OTP
through the web page. The transaction is completed when the
bank receives the valid OTP. However, making banking transac-
tions via the Internet may be subjected to many types of attacks
including password attacks, malware, phishing, and other unau-
thorized activities. The GSM network has several security vulner-
abilities. Only the airway traffic between the mobile station and
the base transceiver station is optionally encrypted with a weak
and broken stream cipher.2�3 The attacker can listen to telephone
conversions and secretly read SMS messages to commit online
crimes.
The SMS OTP is also vulnerable to online phishing attacks.
The attacker may coerce the victim to log into the phishing
site masquerading as the victim’s bank site. The phishing site
∗Author to whom correspondence should be addressed.
captures the username and the password of the victim and
prompts the victim to submit the OTP. Meanwhile, the attacker
uses the victim’s credentials to log into the victim’s bank account
and makes a financial transaction. The bank generates an OTP
and sends it to the victim’s mobile phone. Unaware of the fraud-
ulent activity, the victim enters the OTP and submits it to the
attacker. Subsequently, the attack submits the received OTP to
confirm the transaction to the bank. Since the OTP is identical
to the one sent from the bank, the OTP validation is valid and
the transaction is authenticated. The objective of this paper is to
present transaction authentication schemes for Internet banking
using the mobile OTP and the QR Code. The schemes must be
resilient to known attacks such as eavesdropping, replay, mes-
sage modification, and phishing. The bank sends the transaction
authentication request that is stored in the QR Code and dis-
played on the user’s web browser. The user scans the QR Code,
and the software on the mobile phone calculates the OTP for the
user to submit to the bank through the web browser to confirm
the transaction.
2. BACKGROUND AND RELATED WORKSUser authentication is the process of verifying an identity claimed
by the user.4–6 The user must provide the authentication infor-
mation to prove oneself to a verifying entity. The authentication
information may exist as, or be derived from a knowledge factor
(something the user knows), a possession factor (something the
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3190–3194, 2015
user has), and an inherence factor (something the user is). Two-
factor authentication is an approach, which requires the user to
present two different forms of authentication information.7 The
approach decreases the probability that the user is presenting
false evidence of its identity. Lee et al.8 present an authentication
method using the mobile OTP with the QR Code. The proposed
system relies on the certificate authority to verify the OTP. How-
ever, the proposed authentication system does not protect the user
from online phishing attacks. Comparing the random number that
appears on the computer screen with the one obtained from scan-
ning the QR Code does not add any security since the attacker
can modify both values. The OTP is calculated from the trans-
action information, the perceived time, and the hash of the serial
number of the user’s mobile device. These values are not secret.
Therefore, the generation of the OTP is insecure. Moreover, there
is no explanation how the mobile device reads the transaction
information and the perceived time since these values are not in
the QR Code.
Liao and Lee9 proposed a user authentication scheme based on
the QR Code. In the verification phase, the user sends the ID and
a time stamp to the service provider. The service provider derives
the long-term secret key from hashing the user ID and its long-
term secret key. The service provider sends the QR Code contain-
ing the resulting value from performing a bitwise exclusive-OR
between the long-term secret key and a random number. The
hash of the random number, the user time stamp, and the service
time stamp are also transmitted to the user. For this scheme, an
attacker can pretend to be a valid user and can request for verifi-
cation. Thus, the attacker can obtain many QR Codes. Although
the random number is used only once, the attacker may perform a
bitwise exclusive-OR of the values in the QR Codes to obtain the
exclusive-OR of two random numbers and may be able to learn
a part of the long-term secret key. This is because the long-term
secret key is used many times. It is similar to using one-time
pad or an initialization vector in the cipher block-chaining mode
more than once. Moreover, since many values are not included in
the QR Code, the user has to type them into the mobile phone,
which is inconvenient.
We have also proposed a transaction authentication using
HMAC based OTP10 which can generate the OTP more effi-
ciently. However, the transaction information is sent in clear text.
3. PARALLEL PROVING ALGORITHM
BASED ON SEMI-EXTENSION RULEThe proposed schemes enhance the security of the current Inter-
net banking environment where the user uses username and pass-
word to log into the account. The connection between the client
and the server is done through HTTPS. After the user logs into
the system, a finance transaction can be done by entering the
transaction into the web browser and submitting to the bank’s
server. The bank’s server replies with a QR Code, which contains
relevant information regarding to the transaction, and requests the
user to enter an OTP. The user can use a mobile phone to obtain
the transaction information. The software on the mobile phone
processes the information stored in the QR Code and presents
it to the user. After visually verifying the information, the user
may confirm the transaction by entering the OTP into the web
browser and submitting it. When the bank’s server receives the
OTP, it verifies whether or not the OTP is valid. The transaction
is committed if the OTP is valid and a confirmation is sent to
the user. Otherwise, an error message is displayed on the web
browser. The designs of Internet banking transaction authentica-
tion schemes consist of symmetric cryptography and asymmetric
cryptography is described in the following subsections.
3.1. Symmetric Cryptography Scheme
For this technique, the user must register for the mobile OTP
service in order to use this method as described in Figure 1.
Registering can be done using the ATM. The bank generates a
secret key and sends the QR Code containing the secret key to
the ATM. The user may scan the QR Code to obtain the secret
key. Alternatively, the bank can send print out of the QR Code
via postal mail. This secret key is shared between the bank and
the user. The secret key is saved on the user mobile phone. Using
password-based encryption can protect this secret key. The user
supplies a password, which is used to derive a key to encrypt the
shared secret key. Before using the shared secret key, the user
must enter the correct password. The advantages for this scheme
are the transaction information is encrypted and the encryption
can be performed more quickly than the asymmetric ciphers.
The transaction confirmation process begins when the user
types in the transaction information on the web browser and sub-
mits the transaction information to the bank. After receiving the
transaction request, the bank randomly generates a nonce N1 and
computes the hash value of the transaction information TI and
N1 using a cryptographic hash algorithm. The hash value is used
to verify the integrity of the message. The transaction informa-
tion, which consists of TI, N1, and the hash value, is encrypted
using the shared secret key KA. The encrypted value is encoded
as the QR Code and sent to display on the user’s web browser.
The bank records the OTP issuing time. When the QR Code is
displayed, the user scans the QR Code and decrypts the message
with the shared secret key KA using a mobile phone. The user
computes the hash value of the received TI and N1, and compares
the hash value to the calculated value to verify the integrity of
the received message. If the verification is valid, the user inspects
the transaction information to make sure that it is accurate. The
OTP is derived from N1. The user may type the OTP in the web
browser to proceed with the transaction as illustrated in Figure 2.
Upon receiving the OTP, the bank verifies the OTP’s expiration
time. If the OTP is expired, the transaction is canceled and the
error confirmation is transmitted. Otherwise, the bank proceeds
with the verification. The bank also derives the OTP from N1
using the same technique. Therefore, the bank is able to check
the validity of the received OTP. The valid OTP implies that
the person who has knowledge of the secret key created it. That
User Mobile Phone BankATM
Register for Mobile OTP
Submit Request
Secret Key
QR Code
Secret Key
Select Scan Option
Scan QR Barcode
Fig. 1. Mobile OTP registration for symmetric cryptography scheme.
3191
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3190–3194, 2015
User A Bank
• generate nonce N1
• compute h(TI || N1)
QR[ E(KA, TI || N1 || h( TI || N1)) ]
TI
• KA=secret key
• TI = Transaction Information
• read and decode QR Code• decrypt the message• verify the integrity of the resulting message• compute OTP = f(N1)
OTP
• check if the OTP has not been expired• compute f(N1)• verify if f(N1) is equal to OTP
comfirmation
• KA=secret key
Fig. 2. Transaction confirmation using secret key.
person was able to decrypt the transaction confirmation request
to learn the value of the nonce N1. Encrypting it with the shared
secret key KA protects the nonce N1. Only the user and the bank
know this value. Therefore, the parties with the knowledge of the
secret key can only compute the valid OTP.
3.2. Asymmetric Cryptography Scheme
In order to use this scheme, the mobile OTP registration is
required. However, the registration process for this scheme is
different from the previous two schemes. The user must have a
public-private key pair and the bank’s public key stored on the
mobile phone. The bank must have its own public-private key
pair and the user’s public key. A Certificate Authority (CA) can
be used but it is not required for this scheme to work. This tech-
nique encrypts the transaction information and does not require
shared keys. When the user registers for the mobile OTP ser-
vice, the bank presents the QR Code of its public key certificate
on the ATM. The bank’s public key can be obtained by scan-
ning the presented QR Code using a mobile phone. In Figure 3,
the user public-private key pair is generated at the ATM and is
displayed as a QR Code. The key pair can be transferred to the
mobile phone by scanning the QR Code. The user’s public key
certificate is submitted to the bank via the ATM. Besides using
the aforementioned method, the user may obtain the public key
certificate by other means. For example, the user may use a CA.
However, the key pair must be installed on the mobile phone and
the public key certificate must be submitted to the bank.
When the user desires to make a financial transaction, the user
enters the transaction information and submits it to the bank via
the web browser. The bank randomly generates a nonce N1 and
creates a digital signature on the transaction information and the
nonce using its own private key. For efficiency, a session key
Ks is generated and is used to encrypt signed information which
consists of TI, N1, and the signature. The user’s public key is
used to encrypt the session key. The bank creates the QR Code
User Mobile Phone BankATM
Register for Mobile OTP
Submit Request
Public Key CertificateQR Code
QR Code
Submit User Public Key Certificate
Bank's Public Key
Select Scan Option
Scan QR Barcode
Generate User Key Pair
Click Next
Confirmation
Select Scan Option
Scan QR Barcode
User Key Pair
Register User's Public Key
Fig. 3. Mobile OTP registration for asymmetric cryptography scheme.
of both encrypted entities and sends the QR code to the user’s
web browser. The user obtains the information stored in the QR
Code using the mobile phone. The session key can be obtained
by decrypting the encrypted session key with the user’s private
key. Then, the encrypted information can be decrypted using the
session key. Subsequently, the user verifies the bank’s signature.
If it is valid, the OTP is derived from the nonce N1. Similar to
the two previous schemes, the user must inspect the transaction
information before submitting the OTP to confirm the transac-
tion. The server checks if the OTP is received within the allowing
period. The unexpired OTP is verified by performing the same
computation on N1. The transaction confirmation process for this
scheme is shown in Figure 4. For this technique, the user can
be certain that the received message is from the bank by verify-
ing the digital signature. The bank can be assured that the valid
OTP is from the user who is performing the transaction request
because the encrypted message can only be decrypted with the
private key corresponding to the public key. The user learns the
nonce and uses it to generate the OTP.
4. SECURITY ANALYSESThe proposed schemes use two-factor authentication in which the
password is used to enter the transaction and the key is used to
create an OTP to commit the transaction. We assume that the key
distribution step is done securely. The communication between
the bank and the user’s web browser is secure using HTTPS
protocol. The attacker cannot obtain the information from eaves-
dropping. The eavesdropper cannot break the HTTPS connec-
tion within the limited period. However, the information flow
between the user’s mobile phone and the web browser is inse-
cure. It assumes that the user securely stores the long-term keys,
namely; secret key and private key, on the mobile phone. The
following subsections discuss possible attack types.
4.1. Eavesdropping
Eavesdropping is an unauthorized real-time listening to private
communication. The objective is to obtain information that is
being transmitted. An attacker may be able to capture the QR
3192
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3190–3194, 2015
User A Bank
• generate nonce N1• create signature on TI || N1• generate session key Ks• X1= E(Ks, TI || N1|| signature)• X2= E(ae, Ks)
QR[ X1, X2]
TI
• ae, ad = user A’s key • be = bank’s public key• TI = Transaction Information
• be, bd = bank’s key • ae = user A’s public key
• read and decode QR Code• decrypt X2 using ad• decrypt X1 using Ks• verify signature• compute OTP = f(N1)
OTP
• check if the OTP has notbeen expired
• compute f(N1)• verify if f (N1) is equal to OTP
comfirmation
Fig. 4. Transaction confirmation using public key.
Code from the user’s computer screen. However, the attacker will
not be able to obtain the secret key or the private key informa-
tion since there is no information regarding the key. The attacker
should not be able to decrypt the encrypted information without
the secret key or the private key. Therefore, the valid OTP should
not be generated.
4.2. Message Modification
A message modification attack is an assault on the integrity
of a security system, in which the attacker intercepts mes-
sages, then alters or reorders them to produce an unauthorized
effect. There are two scenarios where the attacker could modify
the transaction information for personal gain. The first scenario
involves modifying the transaction information sent to the bank
and the second scenario involves modifying the QR Code. For
the first scenario, the attacker intercepts the transaction request
The transaction confirmation request(a) (b) The transaction verification result
Fig. 5. Screen captures of the implementation.
and modifies it. For instance, the attacker changes the receiv-
ing account to another account. After transaction information has
been modified, it is submitted to the bank. Correspondingly, the
bank generates the QR code and sends it to the user request-
ing the OTP confirmation. Automatic validation by the software
would be deemed valid. However, when the user inspects the
transaction information, it would be different from the actual
request.
For the second scenario, if the attacker could intercept the
communication and could modify the QR Code, the information
contained in the QR Code would be different from the one sub-
mitted by the user. However, the integrity check would fail. Each
message contains an integrity check using cryptographic hash
code, or digital signature depending on the technique.
4.3. Replay
A replay attack occurs when an attacker repeats a valid data
transmission or delays the original transmission. The aim is to
compromise the integrity of the system. The attacker may be
able to masquerade as the actual party who are making the data
transmission or may cause adverse effects on the system.
All schemes employ the use of the randomly generated number
or the nonce. Therefore, the OTP is unique to the transaction.
Furthermore, the OTP can be used only once within a time limit.
When the user requests a financial transaction, the bank saves
the requested transaction. A nonce and the transaction request
time are also recorded. The transaction is open awaiting the OTP.
If the OTP is received, the transaction is completed. A replayed
confirmation that arrives after the transaction is closed does not
have an effect. On the other hand, a replayed OTP that arrives
during the transaction period does not match the one calculated
by the bank. Hence, the replay attack is infeasible.
4.4. Phishing and Man-in-the-Middle
Phishing is a type of Internet fraud that attempts to acquire a
user’s credentials by deception. The attacker sets up a website
that masquerades as a trustworthy site. Phishing messages are
sent to lure to the user to visit the site. Unknowingly, the user
enters the username and password, which are captured by the
phisher.
Consider the phishing attack scenario where the attacker
entices the user to log into the bank’s site through the attacker’s
site. The attacker may be able to obtain the user’s password.
3193
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3190–3194, 2015
However, the attacker should not be able to make a transaction
since it requires the valid OTP, which was derived from the nonce
using the key, or protected by the key. Suppose that the user
also performs a transaction through the attacker’s website. For
example, the user transfers some money to another account, and
the attacker modifies the transaction information to transfer the
money to the attacker’s account and sends the transaction request
to the bank. The bank sends the QR Code to the attacker and the
attacker presents the QR Code to the user on the computer screen.
When the user scans the QR Code using the mobile phone, the
verification will pass. However, the user will notice that the trans-
action information is different from the one previously requested.
The user will decline to enter the OTP. Without the OTP confir-
mation, the transaction is incomplete. Online phishing is still a
threat since it requires the user to inspect the transaction infor-
mation before entering the OTP. However, a vigilant user will
not be a victim of such an attack.
5. EXPERIMENTAL RESULTSWe implemented the prototypes of the proposed schemes. The
application on the mobile phone is an Android application and
the web application is written in JSP. ZXing library is used for
generating and reading the QR Code. For security API, we use
Java Cryptographic Architecture (JCA). HMAC-SHA256 is used
for generating the message authentication code. AES and RSA
algorithms are used for symmetric and asymmetric cryptographic
operations, respectively. Figure 5 shows the screen capture of
the transaction confirmation request after the user submits the
transaction information. The QR Code contains the transaction
confirmation request, which the user needs to use the key to
create a valid OTP. The OTP for the transaction is valid for three
minutes. Otherwise, the transaction will be canceled. When the
user captures the QR Code using the mobile phone with the
Mobile OTP software, the transaction information is verified and
the OTP is calculated.
The transaction information, the OTP, and the verification
result are displayed as illustrated in Figure 8. The user can visu-
ally verify the transaction information and may submit the OTP
to the bank to commit the transaction.
6. CONCLUSIONSThe transaction authentication using mobile OTP and QR Code
can be implemented in a secure manner. We presented three
authentication schemes namely; symmetric cryptography, and
asymmetric cryptography, that can be securely used to authenti-
cate the user’s transaction. Both schemes can prevent attacks such
as eavesdropping, message modification, replay, and phishing.
For preventing online phishing attacks, the user must visually
inspect the transaction information to ensure its integrity before
submitting the OTP.
References and Notes1. A One-Time Password System, IETF RFC 2289-1998.2. E. Barkan and E. Biham, Conditional estimators: An effective attack on A5/1,
SAC 2005, LNCS, edited by B. Preneel and S. Tavares, Springer, Heidelberg(2006), Vol. 3897, pp. 1–19
3. A. Biryukov, A. Shamir, and D. Wagner, Real time cryptanalysis of A5/1 ona PC, FSE 2000, LNCS, edited by B. Schneier, Springer, Heidelberg, (2001),Vol. 1978, pp. 1–18.
4. Securing mobile devices: Present and future, http://www.mcafee.com/us/resources/reports/rp-securing-mobile-devices.pdf.
5. Internet Security Glossary, IETF RFC 2828-2000.6. W. Stallings, Cryptography and Network Security: Principles and Practice,
6th edn., Prentice Hall, Upper Saddle River, NJ (2013).7. R. J. Boyle and R. R. Panko, Corporate Computer Security, 3rd edn., Prentice
Hall (2012).8. Y. S. Lee, N. H. Kim, H. Lim, H. Jo, and H. J. Lee, Online bank-
ing authentication system using mobile-OTP with QR-code, Proc. 5th Int.Conf. Comput. Sciences and Convergence Information Technology (2010),pp. 644–648.
9. K. Liao and W. Lee, Journal of Networks 5, 937 (2010).10. P. Subpratatsavee and P. Kuacharoen, Transaction authentication using
HMAC-based one-time password and QR code, Computer Science and itsApplications, LNEE (2015).
Received: 9 October 2014. Accepted: 19 November 2014.
Taufiq Immawan1�3�∗, Marimin2�∗, Yandra Arkeman2, and Agus Maulana3
1Department of Industrial Engineering, Faculty of Industrial Technology, Universitas Islam Indonesia2Department of Agroindustrial Technology, Faculty of Agricultural Technology, Bogor Agricultural University, Bogor, 16002, Indonesia
3School of Business and Management, Bogor Agricultural University, Bogor, Indonesia
Batik is an Indonesian cultural asset that should be preserved. In the perspective of its protection, supply chainperformance is important to be assessed, especially in its industrial aspect. Supply Chain Operations Reference(SCOR) is a method of self-assessment and a comparison of activities in the supply chain performance evalua-tion using five attributes; reliability, responsiveness, costs, assets management, and agility. Measurements withSCOR method presents a framework of business processes, performance indicators, best practices, as wellas a unique technology to support communication and collaboration inter-supply chain, so as to improve theeffectiveness of supply chain management and the effectiveness of supply chain improvement. SCOR focusedon causality of the steps that are shaped linear correlation. This state of condition appears to be the weaknessof SCOR method, because it cannot continue more specific steps forward that can predict performance. In thisstudy, SCOR in combination with System Dynamics (SD) can identify the four perspectives and their interactionvariables associated with causal loop design model. The design of a hybrid model SCOR-SD provides moreeffective and interesting value. Furthermore, the simulation of the model from data inputs in 2014 creates predic-tion until 2019. The results obtained for the reliability attributes associated with perfect order fulfillment startedin 2015 to 2019 respectively 85.16% to 91.06%. Responsiveness attributes associated with the order fulfillmentcycle time started at 69.71 days to 62.04 days. The simulation results shown that the associated total cost wasfluctuated ranging from IDR 1,491,041.700 to IDR 1,470,531,891. Attributes associated with cash managementassets to cash cycle time in a row is 35 days to 26 days. The latter attribute is an attribute associated with theagility of supply chain flexibility upside respectively start from 2.26 % to 1.8 days.
Keywords: Supply Chain Operations Reference, Batik Industry, System Dynamics, Performance.
1. INTRODUCTIONMeasurement of the company’s supply chain needs to be done to
keep the distribution of products continuously. Since its introduc-
tion in 1996, the majority of researchers using the supply chain
operations reference (SCOR) to measure the performance of the
supply company.4
There are 5 attributes Introduced in SCOR as a measure of
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3215–3219, 2015
In the USA has conducted a computer-aided supply chain con-
figurations based on the SCOR model. In this study contribute
to the tool-assisted configuration and development2�10 also con-
ducts supply chain performance measures that contribute to the
performance of Chinese furniture manufacturing supply chain.
System Dynamics (SD) is one method that can solve com-
plex problems with simulation models. It is known that the prob-
lems are considered complex systems due to the components
that are inside interact dynamically and provide causation. Such
interaction is called the Causal Loop. Interactions that occur
can be modeled in the form of a mathematical model, which in
turn can be calculated with the help of a computer in the form
of simulation.1 Research using primary approach was made by
Ref. [8], 20075 have conducted an analysis of the value chain by
using a hybrid simulation and AHP. Simulations were carried out,
using an amalgamation of discrete event simulation with con-
tinuous simulation (System Dynamics). However, less obvious
SCOR contribution with SD (System Dynamics) in this study. It
is based on the perspective of the SCOR model is not based on
five attributes in SCOR. According to Ref. [3] this research pre-
sented the development of a dynamic of sustainable supply chain
for Gayo coffee for improving the sustain ability performance.
The dynamic model is proven from the comprehensive descrip-
tion and analysis of the SGCSS systems operation taking
into economic, social and environmental dimension. Economic
dimension focused on actors profits, social dimension empha-
sized on the actors skills and environmental dimension concerned
on the pulp utilization as compost.
The outline of this paper to measure supply chain performance
using a combination of two methods of SCOR and SD, herein
after referred to as Hybrid SCOR and SD to improve dynamic
performance in the batik industry with the MTS-MTO produc-
tion type in Indonesia. Contribution this research are provide the
model for corporate using combination between SCOR perfor-
mance that static measurement and System Dynamic. This com-
bination can be simulated, to predict the final result.
2. LITERATURE REVIEW2.1. Supply Chain Operations Reference (SCOR)
Approach
2.2. SCOR Definition
According to Ref. [4] model of Supply Chain Operations Refer-
ence (SCOR) was delivered by the Supply Chain Council (SCC),
which was established in 1996. The SCOR model created by the
SCC aimed to provide a method of self-assessment and compar-
ison of the activities and performance supply chain as a standard
cross-industry supply chain management. This model presents a
framework of business processes, performance indicators, best
practices (best practices) as well as a unique technology to sup-
port communication and collaboration intra-supply chain, so as
to improve the effectiveness of supply chain management and the
effectiveness of supply chain improvement. Meanwhile, accord-
ing to Ref. [2], SCOR model is a process reference model, which
is intended to make the industry standard that enables supply
chain management in the next generation. Assessing the perfor-
mance parameters such as asset management, cost, reliability,
agility and responsiveness, does evaluation. SCOR performance
section consists of two types of elements: Performance Attributes
and Metrics.4
2.2.1. SCOR Five Performance Attributes
According to Ref. [4] in the supply chain performance measure-
ment through the method of SCOR version 11.0, there are five
attributes of work that will be done to measure the performance,
while the five attributes are:2.2.2.1. Reliability. Reliability is an attribute that is focused
on the consumer. A supply chain should be consumer centric and
companies in the supply chain are need to meet consumer needs.
Reliability stated ability to perform tasks expected. Metric of
reliability consist of right quantity, right in time, right in quality.
Performance indicator is perfect order fulfillment.2.2.2.2. Responsiveness. Responsiveness attribute states how
quickly a task is executed. This shows consistent speed in run-
ning the business. Primary performance indicator is order fulfill-
ment cycle time. Responsiveness focuses on the consumer.2.2.2.3. Agility. Agility attribute states ability to respond to
external changes like the ability to change. External influences
are unexpected demand increase or decrease, stopping operates
from suppliers, disaster, terrorism act, or employee problems.
Primary performance indicator are flexibility and adaptability.
Agility focuses on the consumer.2.2.2.4. Cost. Cost is an attribute that internal focus.
Attributes charge states the cost of running the process. Costs
generally include the cost of labor, raw materials, transportation
costs. Primary performance indicator is total cost to serve. Total
cost focuses on the consumer because of measuring the total cost
to serve the consumer.
2.2.2.5. Asset Management. Asset management attribute states
the ability to efficiently utilize assets. Asset management strategy
in the supply chain include inventory reduction and determine
whether in source or out source productions. Primary perfor-
mance indicator are cash to cash cycle time and return on fix
asset. Asset management focused on internal.
The reason for using 5 attributes are: reliability is used to mea-
sure the level of the fulfillment of order, whether can be right
100 percent or still underneath it. As for the responsiveness is
used to measure the velocity of process level from order accepted
until the order is shipped. The level of responsiveness compare
with the target of the companies. Agility used to measure the flex-
ibility of production, as a reaction over the rise or decline in the
order with 20 percent range. Cost is used to calculate the total
cost to serve so that can be known by the efficiency of the process
of batik production. The value of the total cost to serve compared
with the target company. Asset management used to measure cash
to cash cycle time. Duration time from corporate make payment
for the order, delivering product, until customer cash release.
2.2.2. SCOR Metrics
SCOR metric model includes 134 metric level 1. By using a hier-
archical approach as developed in the process of SCOR, metric
also has several different levels.
(1) Metrics diagnostic level 1 is the overall health of the supply
chain. This metric is also known as strategic metrics and key
performance indicators (KPI). First level of benchmarking met-
ric helps companies set realistic targets to support the strategic
direction.
(2) Metric Level 2 acts as a diagnostic for the metric level 1.
Relationship diagnostic help identify root causes of performance
gaps in the metric level 1.
(3) Metrics level 3 acts as a diagnostic for 2-level metrics.
3216
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3215–3219, 2015
Fig. 1. Causal loop model from five SCOR performance attributes.
2.3. System Dynamics (SD) Approach
SD is a methodology for studying and managing feedback of
variables contained in the complex system1 said the primary
method of studying the problem with a systematic viewpoint,
where the elements of the system interact with each other in
a relationship given feedback to produce a behavior. The inter-
action in this structure is translated into a mathematical model
with the aid of a digital computer simulated to obtain historical
behavior. SD models can create a feedback to decision makers
about the possible absence of the collision of a series of wisdom
to simulate and analyze the behavior of the system on different
assumptions.
Fig. 2. Flow diagram model from five SCOR performance attributes.
2.3.1. Causal/Feedback Loop
Causal loop diagrams or also known as influence diagrams,
are used to help modelers to understand the system by provid-
ing an overview through causality in the system (the system
conceptualization).
Causal loop diagram give simplicity in understanding the
structure and behavior of a system, however the simplicity of it
can also make unclear whether the relationship between variables
happened is the relationship or not rate to level.5 By using the
Causal Loop Diagram modelers can quickly structuring model
based on the assumptions used as in Figure 1. Causal loop dia-
gram used to help model maker understand the system by giving
a general picture via relationship of cause and effect in the
3217
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3215–3219, 2015
Table I. The final result of simulation.
No Attribute Actual Initial simulation Final simulation Target
1 Reliability 85.58% 85.16% 91.06% 100%2 Responsiveness 91.02 days 69.71 days 62.04 days 80 days3 Agility 2.5 days 2.26 days 1.8 days 1.8 days4 Cost IDR 1,482,637,610 IDR 1,491,041,700 IDR 1,470,531,891 IDR 1.400.000.0005 Asset management 32.73 days 35 days 26 days 30 days
system. By causing loop model maker can quickly draw up a
structure model based on assumptions used.5
2.3.2. Flow Diagram
Flow diagram is a representation of a form of detailed depiction
of the system, as in Figure 2. In the flow diagram shown vari-
able types and kinds of relationships between variables in the
system. The main objective of the flow diagram is to represent
the flow and structure of the system in detail in order to facilitate
mathematical modeling.
A diagram describing the relationship between variables made
in the loop diagram case and effect with clear, where used
symbols in certain variables. In the flow diagram distinguished
between the flow of physical and the flow of information. Change
in a variable in this sub system will change physical quantity.
Otherwise the flow of information is not the flow of a convertible.
The information derived from one source could be variable
that transformed to another without reducing the amount of infor-
mation that is in the source.
3. RESEARCH METHODCausal loop that has been created in Figure 1, converted into
flow diagrams and processed using Powersim 9 software. Flow
diagram represent the mathematical model together with the data
processing variables was constructed and described in Figure 2.
4. RESULTS AND DISCUSSIONAttributes associated with perfect reliability orders fulfillment
simulation run for 5 years from 2015 to 2019, following the sim-
89.02%, and 91.06%. From the simulation results can be known
company’s performance can be quite adequate to see the feed-
back that occurs within or between attributes in SCOR. However,
it should be done to improve performance in the future until
100%.
Responsiveness attributes associated with the order fulfillment
cycle time, respectively 69.71 days, 67.37 days, 65.49 days,
64.75 days and 62.04 days. From the simulation, the results
are through the target, are 80 days, better with state of reality,
but the problem of how to minimize the cycle time of service
may be considered for the future. 1,485,517,574, IDR 1, 479,
374,070, IDR Attributes associated with the total cost consecu-
tive associated with the total cost consecutive service charge IDR
1,491,041,700, IDR 1,472, 708, 030 346, and IDR 1.470.531.891.
Attributes associated with cash management assets to cash cycle
time in a row is 35 days, 33.2 days, 31.62 days, 28.66 days, 26
days. The later attribute is an attribute associated with the agility
of supply chain flexibility upside respectively 2.26 days, 2.03
days, 1.98 days, 1.88 days and 1.8 days.
Its better if we use the weighted of each parameter like Analyt-
ical Hierarchy Process (AHP). The systems approach was accom-
plished by identifying all of the factors contained in the system
to obtain a good solution for resolving the problem, and then
creating a model of AHP to help rational decisions. AHP model
is used to calculate the weight of criteria, both quantitative and
qualitative in one research.
Graphically, AHP decision problem can be constructed as
a multilevel diagram (hierarchy). AHP begins with the focus
or goal past the first level criteria, sub criteria, and finally
alternative.6 The final result of simulation can be seen in Table I.
There are some better result between before and after the
simulation. There are some significant improvement because of
lean manufacturing and customer order decoupling point between
Make To Stock and Make To Order.
The managerial implication from this result are the SCOR
parameter hybrid with dynamic model make the model of system
dynamic are able to equip the result of SCOR having the charac-
ter of static. Hybrid SCOR-SD it helps manager of the company
understand interactions among SCOR, of the parameters of so
that by improving one or two of variable will affect the whole of
variable. Manager just looking for variables that have potential
biggest then fix it. The end result is the fifth variable will also
change.
5. CONCLUSION AND SUGGESTION5.1. Conclusion
Reliability for the most influential attribute variable is on sched-
ule production orders (100%), the smallest influence is fulfillment
of orders to customers (58.77%) compared to among the four
main supporting variables.
The most influential attribute for responsiveness is production
cycle time (48.59 days), the smallest influence variable is delivery
time (1.03 days).
The most influential variable in Cost attribute is in the pro-
duction cost and the smallest influence is from procurement vari-
ables. The highest number for asset management is number of
days outstanding loans (57.52 days) and the smallest is number
of variables influence the pending sales (1.03 days).
Meanwhile, the most influential Agility variables is supply
source flexibility (83.33%) and the smallest influence is the make
variable supply flexibility (12.83%).
5.2. Suggestions
It is suggested to apply Analytical Hierarchy Process (AHP) for
weighting the SCOR attributes for measuring the overall supply
chain performance.
Acknowledgment: The Authors would like acknowledged
to School of Business and Management, Bogor Agriculture
3218
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3215–3219, 2015
University, Indonesia and Directorate of Academic Development,
Universitas Islam Indonesia, Indonesia.
References and Notes1. J. W. Forrester, Industrial Dynamics, Massachusetts Institute of Technology,
Massachusetts-U.S.A: The M.I.T. (1961).2. S. H. Huang, S. K. Sheoran, and H. Keskar, Computer and Industrial
Engineering 48, 377 (2005).3. J. R. Machfud and R. S. Marimin, Journal of Theoretical and Applied Infor-
mation Technology 70 (2014).4. J. Paul, Transformasi Rantai Suplai Dengan Model SCOR (Cetakan ke-1
edn.), edited by R. Nurul, A. D. Zalsa, H. Wahyudi, A. Rosyid, and T. Erlinda,Jakarta Pusat, Penerbit PPM (2014).
5. K. E. Maani and R. Y. Cavana, System Thinking and Modelling UnderstandingChange and Complexity, Prentice Hall, New Zealand (2000).
6. D. M. A. Marimin, P. M. P. I. F. P. Machfud, and B. Wiguna, Journal of CleanerProduction 85, 201 (2014).
7. J. A. Palma-Mendoza, International Journal of Information Management 34,634 (2012).
8. F. Persson, International Journal of Production Economics 131, 288(2011).
9. L. Rabelo, H. Eskandari, T. Shaalan, and M. Helal, International Journal ofProduction Economics 105, 536 (2007).
10. D. J. Robb, X. Bin, and T. Arthanari, International Journal of Production Eco-nomics 112, 683 (2008).
11. V. Salton and S. H. Belton, Adding Value to Performance Measurementby Using System Dynamics and Multicriteria, Strathcyde Business School,Research Paper No. 2001/19 (2001).
Received: 23 October 2014. Accepted: 15 December 2014.
Department of Industrial Engineering, Faculty of Industrial Technology,Universitas Islam Indonesia, Yogyakarta, Indonesia
A practical and novel approach for probabilistic inventory analysis is presented in this paper. In uncertaintyenvironment, where demand of every period is not constant, then the risk to have stock out can be occurredduring ordering lead time. In this study, since the inventory can be monitored every time, hence, continuousreview technique is proposed as the basis of analysis. The nature of continuous probabilistic inventory analysisis very complicated. Even though it can minimise inventory cost, however, such technique is not practical forindustrial world due to its complication. To model probabilistic factors, besides mathematical model, Fuzzy Logic(FL) can be used for that purpose. In FL, probabilistic factors that cause uncertainties will be represented usingFuzzy sets. However, in conventional FL, the Fuzzy sets are developed using precise curves and sometime itcannot cope with uncertainties. Therefore, in this study, a Type-2 FL is proposed to model probabilistic factorsin the inventory analysis. A case study shows that the proposed model able to give reasonable solution to theproblem.
index, �x=membership value of input x, ↓= lower Fuzzy curve,
↑= upper Fuzzy curve.
To evaluate how the proposed model works, an input vector is
required. Let say, i = �0�60�0�58�0�48�0�63� is the input vector,
then the result of Fuzzy rules evaluation is as shown in Table II.
By applying Karnik-Mendel algorithms, shifting points are
obtained, with upper to lower (ul) point is 8 and lower to upper
(lu) is 8. Hence, output from ul and lu can be calculated as
follows:
yul =�∑8
i=1 fi ↓×yi ↑�+ �∑15
i=9 fi ↑×yi ↑�∑8i=1 fi ↓+
∑15i=9 fi ↑
(5)
yul = 1
ylu =�∑8
i=1 fi ↑×yi ↓�+ �∑15
i=9 fi ↓×yi ↓�∑8i=1 fi ↑+
∑15i=9 fi ↓
(6)
ylu = 0�4
Table I. The fuzzy rules.
Rule D Dl h A Q
R1 H H H H M (0.4–0.6)R2 H M N H H (0.7–1.0)R3 H M N N M (0.4–0.6)R4 M H H H M (0.4–0.6)R5 L M H N L (0.2–0.4)R6 L L H N L (0.2–0.4)R7 L M N N M (0.4–0.6)R8 M M N N M (0.4–0.6)R9 H L N H L (0.2–0.4)R10 H M N H H (0.7–1.0)R11 M H H N M (0.4–0.6)R12 L L N N L (0.2–0.4)R13 H L H N L (0.2–0.4)R14 M L H H M (0.4–0.6)R15 L M N H M (0.4–0.6)
Note: H = high, M =medium, L= low, N = normal.
3242
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3240–3243, 2015
The Q value above is still normalised value. When it is con-
verted to the real Q value, then 120 units is obtained. The r value
can be calculated as follows. Assuming � = 500.∫ Dl
r
1
Dldx = hQ
�D(8)
∫ 25
r
1
25dx = 115×120
500×100
r
25= 1−0�276
r = 18
And E�x� = −5�9 units. Since E�x� is less than 0 then back
order will not occurred. Total cost for the decision proposed by
T2FL is as follows.
TC�120�18� = 225�000×100
120+115
(120
2+18−25
)+0
TC�120�18� = 193�593
4. DISCUSSIONFuzzy logic usually is developed by experts. The proposed T2FL
is also developed by the experts who have wide experiences about
the investigated inventory system. Therefore, the result can be
considered as feasible solution for the investigated system. How-
ever, T2FL is not an optimisation tool, the result provided by the
proposed T2FL may be not the best solution. Total cost of the
solution may be not the most minimum one.
The advantage of the proposed T2FL is the ability to cope with
uncertainty of input variables. In optimisation model of inventory
system, which is based on mathematical model, such uncertainty
is not responded. In real industrial applications, uncertainty will
frequently occurred and T2FL will be more applicable compared
to the mathematical model.
5. CONCLUSION AND SUGGESTIONIn T2FL, fuzzy set for every input variable has an interval. In
the consequent part, output also modelled in the form of interval.
Such condition is believed can cope with uncertainties. Based on
the explanation above, it is proven that the proposed T2FL can be
used to model continuous review probabilistic inventory system
with backorder case.
For the next study, it is proposed to hybrid the proposed T2FL
with an optimisation algorithm such as Genetic Algorithms (GA).
The optimisation tool can be used to adjust parameter values
inside the T2FL so that an optimum solution can be obtained.
References and Notes1. E. A. Elsayed and T. O. Boucher, Analysis and Control of Production System,
Prentice-Hall, Inc., New Jersey (1994).2. J. M. Mendel, Uncertain Rule-Based Fuzzy Logic Systems: Introduction and
New Directions, Prentice-Hall, Upper Saddle River, NJ (2001).3. S. H. R. Pasandideh, S. T. A. Niaki, and N. Tokhmehchi, Expert Systems with
Applications 38, 11708 (2011).4. I. Saracoglu, S. Topaloglu, and T. Keskinturk, Expert Systems with Applications
41, 8189 (2014).5. L. Wang, Q.-L. Fu, and Y.-R. Zeng, Expert Systems with Applications 39, 4181
(2012).6. O. Mendoza, P. Melín, and O. Castillo, Applied Soft Computing 9, 1377 (2009).7. S. Miller and R. John, Knowledge-Based Systems 23, 363 (2010).8. M. H. F. Zarandi, M. R. Faraji, and M. Karbasian, Applied Soft Computing
12, 291 (2012).
Received: 30 November 2014. Accepted: 28 January 2015.
Malik Bader. Alazzam∗, Abd. Samad Hasan Basari, Abdul Samad Sibghatullah, Mohamed Doheir,Noorayisahbe Mohd Yaacob, and Farah Aris
Faculty of Information and Communication Technology, Universiti Teknikal Malaysia Melaka (UTeM),76100 Durian Tunggal, Melaka, Malaysia
Electronic health records (EHRs) exchange improves hospital quality and reduces medical costs. However, fewstudies address the antecedent factors of physicians’ intentions to use EHR exchange. Based on institutionaltrust integrated with UTAUT2 model, we propose a theoretical model to explain the intention of physicians touse an EHR exchange.
liness; (e) Efficiency; (f) And equity.7 Without a mechanism to
exchange EMR among hospitals, the inability of hospital staff
to review the medical history of patients who have visited other
hospitals could also result in redundant.4
However, the implementation of an EMR exchange system
requires considerable effort. Prior research has identified cer-
tain factors of EMR exchange adoption and implementation
among hospitals,5 use experiences regarding health information
exchange (HIE),6 the clinical application of HIE,7 HIE-related
privacy and security concerns], and public attitudes toward HIE.8
It is critical to encourage physicians to use the EMR exchange
system. Despite this urgent need, scant studies have explored
the antecedent factors of physicians’ intentions to use EHR
exchange. In a typical medical setting, the characteristics of the
users, technology, norms, and context may substantially differ
from those in business settings; thus, this study explored the
antecedent factors of physicians’ intentions to use EMR exchange
from the perspective of information system adoption.
1.3. UTAUT Model Review
1.3.1. UTAUT1
Reference [9] academics have showed technology acceptance
studies for over two decades now. They have used numerous con-
cepts and models to carry out these papers in different contexts
with different part of study. Findings from these researches vary.
The authors of UTAUT model unified eight concepts and mod-
els which contain Theory of reason Action (TRA) Technology
acceptance model (TAM), Motivational model (MM), Theory of
planned behavior (TPB) combined TAM and TPB (C-TAM-TPB)
Model of PC Utilization MPCU. Innovation Diffusion Theory
(IDT) and Social Cognitive.9�10
Theory (SCT) Bandura (1986). The unification by the inves-
tigators sum up all the concepts from the eight models to four
elements, which expects intentions, usage, and four moderators
of the key relationships.11 Figure 1 explains the relationships that
exist in the UTAUT model. The model has four EV, which refers
to exogenous variables, EE, which refers to effort expectancy,
PE which indicates to performance expectancy, SI which refers
to social influence, and FC which mea facilitating conditions.
The endogenous variables are the technology intention to use
and behavior. There are other four moderators age, namely, gen-
der, experience, and voluntariness. Performance expectancy is
famous, as a degree individual believes in the benefit of the sys-
tem to performance.9�12�13
The degree of ease linked with the use of the system is an
important indicator towards technology intention to use which
calls effort expectancy. The degree of an individual perceives on
the important of new system used is also significant indicator
towards technology intention to use. The degree of an individual
believes on the effective of organizational and technical infras-
tructure exists that needs to support the use of the system is an
important indicator which called facilitating condition.
The chosen of this model in this paper is justified by its univer-
sal and integrative way, incorporating a vast variety of explanatory
variables from the main theoretical models developed to illustrate
technology acceptance and use. In particular, Morris et al.11 car-
ried out an in-depth test of literature on this matter and proposed
a unified model that merge the contributions collective to the pre-
vious theories. Therefore, it is reasonable to expect a theory that
integrates the most important contributions from other models to
be more to the previous theories explanation of technology accep-
tance and use.
1.3.2. UTAUT2
Reference [14] extends the unified theory of acceptance and use
of technology (UTAUT) to investigate acceptance and use of
technology in a consumer context. That the goals of UTAUT2
integrates three concepts into UTAUT: HM, PV, and HT the
demographic characteristics of service users’ were used as mod-
eratos variables namely experience, age and gender to control the
effect on the BI and the use of technology. The findings have
derived from two-channels online survey conducted with user of
technology. The data collected took four months from 1,512 of
the clients of mobile. As compared to UTAUT, the additions pro-
posed in UTAUT2 produced a substantial improvement in the
variance explained in BI.
3250
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3249–3253, 2015
Fig. 1. UTAUT2 conceptual model.14
1.4. Hedonic Motivation (HM)
Hedonic motivation (HM) can defined as the intrinsic motivation
such as fun, enjoyment or pleasure when using a technology
because of technology for its own sake, and it has known an
important construct in determining technology acceptance and
use. HM is similar to perceived enjoyment or playfulness to TAM
as an intrinsic motivation factor.14–16
1.5. Price Value (PV)
In general, people chose the services or products when their
benefit gives more than the price value compared with its cost.
Therefore, price value can be defined as learners’ cognitive trade-
off among the seeming benefits of the applications and the eco-
nomic cost.14�16�17
1.6. Habit (HT)
Habit (HT) is one of a strong predictor of future technology
use.17 Habit has been identified as the degree to which individuals
incline to implement behaviors routinely due of learning.9�14�15
1.7. Institutional Trust
Physicians have limited ability to monitor or control the EEC
use of their EHR exchange, which is why trust is required.
Institution-based trust exists when trust is associated with the
existence of third-party structures that are independent of dyadic
actions. Shapiro41 described institutional trust as the belief that a
trustor has regarding the security of a situation because of guar-
anteed safety procedures and other structures.1 Zucker indicated
that institutional trust is the most crucial trust-creating mode
among impersonal economic environments, where a sense of a
community with similar values is lacking.
Institutional trust measures were employed to address physi-
cian perceptions regarding whether the national framework
(Legal and regulatory framework, Third-party guarantees, Inter-
national standards, Directives, Escrows) was conducive to using
an EHR exchange system. Regarding technology, the EEC facil-
itates using e-signatures to replace physical signatures or seals
on medical records in accordance with the law for verifying the
identity of a signee. EHR security can be enhanced using certified
encryptions to protect the privacy of patients and the integrity
of medical records, and requiring certified decryptions to access
records.1
Regarding legality, the DOH continues to amend its regulations
of EHR production and management for medical institutions; this
regulation promotes self-management and continual improvement
among EHR exchanges. Concerning policy, the DOH formulates
EHR inspection mechanisms, using the International Organiza-
tion for Standardization (ISO) 27001 guidelines to formulate
inspection items in accordance with the regulation of EHR pro-
duction and management by medical institutions. The EEC pro-
vides certified EHR logos, enabling members of the public to
identify which hospitals have adopted certified EHR and pro-
moting EHR exchanges among hospitals. Regarding standards,
the EEC provides standardized templates based on the specifica-
tions of EHR exchanges, ensuring the appropriate EHR content
to minimize infringement risks to patient privacy. Therefore, in
this study, institutional trust is a critical research
1.8. Research Methodology
1.8.1. EHRs and UTAUT2
The UTAUT model has been widely used in the EHR adoption
and acceptance which is shown in Table I. The employees will
find the EHR system (physicians) useful if it helps them to per-
form the functions of the Directorate efficiently and effectively.
PE, EE, SI, HD, PV (price value has been excluded due deferent
related this study area), HB will directly affect the intention to
use of the EHRs by the officers and staff. Thus, a high level of
intention to use is likely to increase employee adoption of EHRs.
H1. Performance expectancy is positively related to physician
intention to use EHR system.
H2. Effort expectancy is positively related to physician intention
to use EHR system.
H3. Social influence is positively related to the physician inten-
tion to use EHR system.
H4. Facilitating conditions is positively related to physician the
intention to use EHR system.
3251
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3249–3253, 2015
Fig. 2.
H5. Hedonic motivation is positively related to physician inten-
tion to use EHR system.
H6. Habit is positively related to physician intention to use EHR
system.
H7. trust institutional is positively related to physician intention
to use EHRs.
2. FUTURE RESEARCH AND LIMITATIONSThis paper is in its first hypothetical idea, in which a beginning
model is proposed based on the previous studies and theoretical
thought. The following phase is the application and validation
of the model to a arrangement of healthcare professionals, in
order that to test the established framework and directly measure
its explanatory and predictive power. Coming studies may eval-
uate other relationships that were not expected in this model and
that will develop the ability to describe the dependent variables.
Therefore, this paper opens up other selections for future research
Refinement of the constructs and measures is one of the possi-
bilities. Additional option is the examination of more complex
relationships between the Independent and dependent variables
of the model. Testing this model with other e-health technolo-
gies, and in other countries that may be more or less developed
than developed countries in e-health use are options that can also
bring benefit.
3. CONCLUSIONSUnderstanding the acceptance and use of EHR system of physi-
cians should bring strong benefits for the future sustainability of
the Healthcare System, which will enjoy more efficient use of
resources. Therefore, the goal of this paper is to detect a set of
determinants of acceptance of EHRs by physicians. To realize
this goal, we suggest a research model based on UTAUT2, adding
new constructs trust institution. We designate this new set of con-
structs “e-health extension to UTAUT2.” We also suppose this
paper to provide a theoretical framework that is a foundation and
a starting point for future research on the acceptance of EHRs
by physicians.18
Acknowledgment: This paper is part of Doctor of philoso-
phy (Ph.D.) work in UteM and part of the work also supported
under grand PJP/2013/FTMK (17A)/S01230.
References and Notes1. P.-J. Hsieh, Int. J. Med. Inform. 84, 1 (2015).2. M.-P. Gagnon, E. K. Ghandour, P. K. Talla, D. Simonyan, G. Godin,
M. Labrecque, M. Ouimet, and M. Rousseau, J. Biomed. Inform. 48, 17(2014).
3. K. Jammoul, H. Lee, and K. Lane, Understanding Users Trust and the Mod-erating Influence of Privacy and Security Concerns for Mobile Banking: anElaboration (2014), Vol. 2014, pp. 1–11.
4. M. N. Herian, N. C. Shank, and T. L. Abdel-Monem, Health Expect. 17, 784(2014).
5. C. Kanger, Evaluating the Reliability of EHR-Generated Clinical OutcomesReports: A Case Study (2014), Vol. 2.
6. S. Trang and S. Zander, Dimensions of Trust in the Acceptance of Inter–Organizational Information Systems in Networks: Towards a Socio-TechnicalPerspective (2014).
7. J. F. Cohen, Trust, Risk Barriers and Health Beliefs in Consumer Acceptanceof Online Health Services (2014), pp. 1–19.
8. A. Balaid, M. Z. A. Rozan, and S. N. Abdullah, Asian Soc. Sci. 10, 118 (2014).9. M. D. M. B. Alazzam, Dr. A. S. Sibghatullah, and A. S. H. Basari, Eur. Sci. J.
10, 249 (2015).
3252
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3249–3253, 2015
10. A. A. Taiwo and A. G. Downe, The Theory of User Acceptance and Use ofTechnology (UTAUT): A Meta-Analytic Review of Empirical Findings (2013),Vol. 49.
11. M. G. Morris, M. Hall, G. B. Davis, F. D. Davis, and S. M. Walton, User accep-tance of Information Technology: Toward Aunified View 1 (2003), Vol. 27,pp. 425–478.
12. M. J. Wills, Examining Healthcare Professionals Acceptance of ElectronicMedical Records Using Utaut (2008), Vol. IX, pp. 396–401.
13. A. Hennington and B. D. Janz, Information Systems and Healthcare XVI:Physician Adoption of Electronic Medical Records: Applying the UTAUT Modelin a Healthcare Context Records: Applying the UTAUT Model in A (2007),Vol. 19.
14. V. Venkatesh, Consumer Acceptance and Use of Information Technology:Extending the Unified Theory (2012), Vol. 36. pp. 157–178.
15. M. Kang, B. T. Liew, H. Lim, J. Jang, and S. Lee, Emerging Issues in SmartLearning (2015), pp. 209–216.
16. M. Technologies, S. D. Impact, and M. I. Usage, Mobile Technologies andServices Development Impact on Mobile Internet Usage in Latvia MobileTechnologies and Services Development Impact on Mobile Internet Usage inMobile Technologies and Services Development Impact on Mobile InternetUsage in Latvia (2013).
17. A. Raman and Y. Don, Int. Educ. Stud. 6 (2013).18. J. Tavares and T. Oliveira, Electronic Health Record Portal Adoption by Health
Care Consumers Proposal of a New Adoption Model (2013), pp. 387–393.
Received: 17 January 2015. Accepted: 20 February 2015.
Guruh Fajar Shidik∗, Syafiq Wardani Dausat, Rima Dias Ramadhani, and Fajrian Nur Adnan
Dept. Computer Science, Universitas Dian Nuswantoro, Semarang 50131, Indonesia
Biometric identification is a security method that utilize parts of human body. Finger Knuckle Print (FKP) is oneof many parts in human that relatively unique. This research try to find out the best distance measurementas recognition method with implication of pre-processing CLAHE. Several variants of distance measurementthat evaluated in this research are Euclidean, Manhattan, Minkowski, Canberra, Chi-square, Chebyshev andBray Curtis. Besides that, feature extraction Principal Component Analysis (PCA) was applied as texture featureafter preprocessing phase. The experiment results has shown several recognition that used pre-processingCLAHE able to improve accuracy with distance measurements in recognize FKP, moreover, several distanceshowed degraded accuracy. Highest accuracy that used pre-processing CLAHE are gathered when applied withChi-Square distance until 95.15%.
In 100-feature PCA, Euclidean distance, has accuracy until
89.69%, Manhattan distance as much as 88.48%, Minkowski
distance as much as 87.87%, Chebyshev distance as much as
80,60%, Chi square distance as much as 87.87%, Canberra dis-
tance as much as 76.96% dan Bray Curtis distance until 89.69%.
In 120-feature PCA, Euclidean distance has accuracy until
89.69%, Manhattan distance until 88.48%, Minkowski distance
as much as 87.87%, Chebyshev distance as much as 80.60%, Chi
square distance sebesar 86.06%, Canberra distance as much as
79.39% and Bray Curtis distance as much as 88.48%.
The results in the Table I, obtained from the pre-processing
without using CLAHE, show that by using 60-reduction, the
Euclidean Distance will have the accuracy value of 90.30%,
Manhattan distance as much as 89.09%, Minkowski distance as
much as 87.27%, Chebyshev distance as much as 80.60%, Chi
square distance as much as 88.48%, Canberra distance as much
as 76.36% and Bray Curtis distance as much as 89.69%.
In the Euclidean distance 100-reduction, it has the accu-
racy value of 89.69%, Manhattan distance as much as 88.48%,
Minkowski distance as much as 87.87%, Chebyshev distance
as much as 80,60%, Chi square distance as much as 87.87%,
Canberra distance as much as 76.96% dan Bray Curtis distance
sebesar 89.69%.
In the Euclidean distance 120-reduction, it has the accu-
racy value of 89.69%, Manhattan distance sebesar 88.48%,
Minkowski distance as much as 87.87%, Chebyshev distance as
much as 80.60%, Chi square distance sebesar 86.06%, Canberra
distance as much as 79.39% and Bray Curtis distance as much
as 88.48%.
Results in the Table II, obtained from the pre-processing using
CLAHE, show that by using 60-reduction, the Euclidean Dis-
tance will have the accuracy value of 80.60%, Manhattan distance
as much as 93.33%, Minkowski distance as much as 70.90%,
Chebyshev distance as much as 64.84%, Chi square distance as
88.68 87.67
Average of Accuracy FKP Recognition without Pre-processing CLAHE
87.47
77.57
89.29
80.60
89.8995.00
90.00
85.00
80.00
75.00
70.00
Euclide..
.
Manha..
.
Minko
...
Chebyc
...
Chi...
Canber
...
Bray..
.
Fig. 2. The accuracy average of FKP without CLAHE.
3277
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3275–3278, 2015
Average of Accuracy FKP Recognition withPre-processing CLAHE
80.80
94.54
120.00
80.00
100.00
60.00
20.00
0.00
40.00
Euclid...
Manha..
.
Minko
...
Cheby..
.
Chi...
Canbe..
.
Bray..
.
70.90 64.84
95.35
86.46
94.14
Fig. 3. The accuracy average of FKP using CLAHE.
much as 95.15%, Canberra distance as much as 89.69% and Bray
Curtis distance as much as 92.72%.
In Euclidean Distance 100-reduction, it has the accuracy value
of 80.60%, Manhattan distance as much as 94.54%, Minkowski
distance as much as 70.90%, Chebyshev distance as much as
64.84%, Chi square distance as much as 95.15%, Canberra dis-
tance as much as 85.45% and Bray Curtis distance as much as
94.54%.
In Euclidean Distance 120-reduction, it has the accuracy value
of 81.21%, Manhattan distance as much as 95.75%, Minkowski
distance as much as 70.90%, Chebyshev distance as much as
64.84%, Chi square distance as much as 95.75%, Canberra dis-
tance as much as 84.24% and Bray Curtis distance as much as
95.15%.
In the other hands, accuracy of Euclidean, Minkowski and
Chebyshev distance are decreased when applied pre-processing
CLAHE. This conditions are showed the uses of pre-processing
CLAHE not always gives significant implication even quality of
images has been improved.
The experiment results in the Figure 2 are the values from the
accuracy average without CLAHE, showing that the Euclidean
distance has the accuracy value of 89.89%, Manhattan distance
as much as 88.68%, Minkowski distance as much as 87.67%,
Chebyshev distance as much as 80.6%, Chi square distance as
much as 87.47%, Canberra distance as much as 77.57% and Bray
Curtis distance as much as 89.28%. Thus, in the pre-processing
step without using CLAHE, the highest accuracy is showed in
the distance measurement using Euclidean Distance as much as
89.89%.
Figure 3 shows the experiment results based on average values
from the pre-processing accuracy of Finger Knuckle Print using
CLAHE by using the distance measurement in the 60, 100 and
120 reductions.
The experiment results in the Figure 3 are the values from
the accuracy average using CLAHE, showing that the Euclidean
distance has the accuracy value of 80.80%, Manhattan distance
as much as 95.54%, Minkowski distance as much as 70.90%,
Chebyshev distance as much as 64.84%, Chi square distance as
much as 95.35%, Canberra distance as much as 86,46% and Bray
Curtis distance as much as 94.14%. Thus, in the pre-processing
step without using CLAHE, the highest accuracy is showed in
the distance measurement using Euclidean Distance as much as
89.89%.
Figures 2 and 3 are calculate the total average accuracy with
feature extraction PCA 60, 100, and 120 in all distance measure-
ment. In Figure 2 have showed Euclidean as Best distance mea-
surements technique that has average accuracy until 89.89% in
recognition FKP without pre-processing. The experiment results
in the Figure 3 are showed Chi square distance measurement has
highest average accuracy until 95.35% with improvement accu-
racy until 8.2% compared with non-preprocessing.
7. CONCLUSIONThis experiment shows the use of different distance measurement
highly influences the accuracy of recognition. Moreover, the uses
of pre-processing CLAHE to improve image quality has shown
significant improvement in several distance measurements such
as Manhatan, Chi-Square, Canberra and Bray Curtis. However,
the implication of CLAHE in some distance gives worst result
that caused distance measurement such as Euclidean, Minkowski,
and Chebychev are decreased.
The used of Chi-Square distance are considered as best dis-
tance measurements that have best accuracy in recognizing
Finger Knuckle Print with pre-processing CLAHE that showed
accuracy until 95.35%.
References and Notes1. K. Usha and M. Ezhilarasan, Computers and Electrical Engineering
(2014).2. L. Zhang, L. Zhang, D. Zhang, and H. Zhu, Pattern Recognition 43, 2560
(2010).3. Z. S. Shariatmadar and K. Faez, Optik—International Journal for Light and
Electron Optics 125, 908 (2014).4. S. Ribaric and I. A. Fratric, IEEE Trans. Pattern Anal. Mach. Intell. 27, 1698
(2005).5. A. K. Jain, A. Ross, and S. Pankanti, A prototype hand geometry based
verification system, Proceeding AVBPA, Washington, DC (1999), Vol. 1,pp. 166–71.
6. Z. Lin, Z. Lei, and Z. David, Finger-Knuckle-print: A new biometric identifier,Proceeding IEEE International Conference on Image Processing, Cairo, Egypt(2009), Vol. 1, pp. 76–82.
7. D. L. Woodard and P. L. Flynn, Comput. Vis. Image Underst. 100, 357 (2005).8. N. Ozkaya and N. Kurat, Journal of Visual Communication and Image Repre-
sentation 25, 1647 (2014).9. Z. Lin, Z. Lei, and Z. David, Finger-Knuckle-Print Verification Based on
10. Y. Pengfei, Z. Hao, and L. H. Yan, Appl. Mech. Mater. 44, 703 (2014).11. C. Ravikanth and A. Kumar, Biometric authentication using finger-back sur-
face, Proceedings of the CVPR’07 (2007), pp. 1–6.12. A. Kumar and C. Ravikanth, IEEE Transactions. Information Forensics and
Security 4, 98 (2009).13. G. Suprijanto, J. E. Azhari, and L. Epsilawati, Image contrast enhancement for
film-based dental panoramic radiography, International Conference on SystemEngineering and Technology (2012).
14. S. M. Pizer, E. P. Amburn, J. D. Austin, R. Cromartie, A. Geselowitz, T. Greer,B. T. H. Romeny, J. B. Zimmerman, and K. Zuiderveld, Computer VisionGraphics and Image Processing 39, 355 (1987).
15. M. Turk and A. Pentland, J. Cognitive Neuroscience 3, 71 (1991).16. S. Szabolcs, Color histogram features based image classification in content-
based image retrieval systems, 6th International Symposium on AppliedMachine Intelligence and Informatics (2008), pp. 221–224.
17. D. C. Adams, F. J. Rohlf, and E. S. Dennis, Italian Journal of Zoology 71, 5(2004).
18. V. Asha, GLCM Based Chi-Square Histogram Distance for Automatic Detec-tion of Defects on Patterned Textures.
Received: 8 October 2014. Accepted: 16 November 2014.
Faculty of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Bangkok 10520, Thailand
The distortion of ultra wideband impulse radio (UWBIR) system that distorted by a channel due to antennadispersion. This highly degrades of link budget performance. Therefore, to know the antenna characteristics,the effects of a waveform distortion, are necessary. The UWB transmission waveform is investigated by usingextended Friis transmission formula based on UWB measurement data. This paper evaluates the performancesystem on transmission waveform distortion for UWB communication. The received signal and isotropic signaltemplates are considered with Friis formula. Mostly of radio wave propagation in link budget analytic performanceis evaluated by using the Friis equation. Due to Friis equation is cannot directly apply for the UWB transmis-sion waveform. For experimental evaluation scheme, using the broadband antenna for transmitter and receiverantennas (Tx and Rx). The full band 3.1 GHz to 10.6 GHz, which contributed by the Federal CommunicationsCommission (FCC), is proposed as the UWB signal waveform. The transfer functions measured as experimentalresult by using the vector network analyzer for measuring and recording. This paper UWB transmission gainswith the received signal and isotropic signal templates are shown and compared.
Keywords: UWB, UWB Antenna, Link Budget, Distortion.
1. INTRODUCTIONThe nowadays UWB system is new for wireless system has
become important for short range wireless communications sys-
tem with its high speed communication, low power consumption
potentials and low.1�2 Therefore, the technology of UWB is dif-
ferent from other technologies. The UWB radio transmits is with
impulse waveform and wider bandwidth of the norrowband spec-
trum. The FCC regulation provides that UWB spectrum from
3.1 GHz–10.6 GHz,13 and also has been greater than 0.20 of the
fractional bandwidth. The UWB fractional bandwidth defined as
BWf = 2�fH −fL�
fH −fL≥ 0�20 (1)
where fL is minimum and fH is maximum of frequency.
UWBIR transmission waveform under the FCC, with part
15 limits has power spectral density under −41.3 dBm/MHz,
which taken as a lower noise floor. Thus, the reason why
UWB communication system can coexist within other microwave
communications and propagation engineering. Furthermore, the
UWBIR system is an ideal trend that new communication tech-
nology for wireless communication, short impulse radio system,
high speed transmission rate and low cost technology for indoor
systems in covered wireless personal area networks (WPAN) and
wireless body area network (WBAN).3
∗Author to whom correspondence should be addressed.
In the communication systems has used Friis transmission for-
mula and it is widely used for calculating the propagation chan-
nel for narrowband communications.4 The complex form in the
free space of Friis formula expression is modified for UWBIR
system.5–7 The UWBIR system, conceded used matched filter.8–10
Although, the frequency spectrum and signal distorted by energy
channel transfer function is used for deriving the SNR gains,11
if considerations about the measured frequency transfer function
and UWB antenna transfer function.
The performance evaluation of waveform distortion due to
channels and antennas for UWB impulse radio. The UWB
receiver considering using received signal and isotropic signal
templates provided to analyze of noise level between input wave-
form and output waveform is obtained. Herein, both biconical
antennas are considered both at transmitter (Tx) and receiver
(Rx) sides. Full band spectrum frequency, which is followed FCC
both indoor and outdoor spectral mask, provided as the UWB
transmitted.12�13 The UWB transmitted is measured and recorded
by using a VNA work done in an anechoic chamber. The band-
width of measurement is from 3–11 GHz. The waveform distor-
tions are considered for the UWBIR transmission waveform.
The transmission gain model of UWB impulse radio receiver
with received signal and isotropic signal templates is shown and
compared with the experimental results and conclusion. This
scheme provides some useful physical insights and optimized
design procedure with a clear and accessible description of the
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3289–3292, 2015
In this organized of paper as follows. Section 2 the UWB
impulse radio measurement system. Section 3 the distortion anal-
ysis of UWB impulse radio system. Next, the experimental
results and discussion are illustrated in Section 4. Finally is con-
clusion in section 5.
2. UWB IMPULSE RADIO MEASUREMENT
SYSTEM2.1. UWB Waveform Model
The transmission waveform is more obvious in the UWBIR sys-
tem. In this paper, the transmitted waveforms that fully following
FCC from 3.1 GHz to 10.6 GHz13 and common frequency band
FCC in USA are considered. Also, the root raised cosine (RRC)
model is used as the waveform transmitted.
The waveform as passband in root raised cosine is written as
Vt�ro�f �=
⎧⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎩
1 �f �−fc � ≤�1−��
2T
A�1−��
2T< ��f �−fc � ≤ �1+��
2T
0 otherwise
(2)
where
A=√
1
2
[1+cos
(�T
�
[��f �−fc �−
1−�
2T
])](3)
T = 1/fb denotes the reciprocal of the symbol-rate and fb is
the spectral bandwidth, fc denotes the center frequency, �= 0�3satisfies roll-off factor. For following FCC spectral masks, fc was
6.85 GHz. The spectral bandwidth fb was 6.37 GHz. For follow-
ing the common frequency band spectral mask, then fc and fbwere 7.877 and 0.975 GHz respectively.
The power spectral density normalized of these waveforms sat-
isfying FCC spectral masks shown in Figure 1.
2.2. Experimental System
The UWB waveform distortion is experimented are considered
used VNA for measuring and recoding in free space. The VNA
Fig. 1. The power spectral density with root raised cosine pass band wave-form by satisfying FCC spectral mask.13
operates in the channel sounding mode, where Two-ports are Tx
antenna and Rx antenna respectively. Both Tx and Rx antennas
height at 1.75 m and separated by 4 m. This experimental setup
is shown in Figure 2 and antenna orientation was rotated Rx
antenna rotation from 0� to 360� by step 5� and antenna polar-
ization in horizontal is measured.
In the calibration techniques of UWB transmission link was
conducted by back to back transmission and short open load
thru (SOLT). Therefore, all impairments of characteristics of the
biconical antennas are verified with the measured results.
3. DISTORTION ANALYSIS OF UWB SYSTEM3.1. Transmission Waveform
The UWB transmission waveform, free space link budget is
transformed by a transfer function. Thus, in free space transfer
function Hf �f � is written as
Hf �f �d� = c
4�fde−j2�fd/c (3)
The transfer function in free space Hc�f � with the biconical
antennas is obtained based on modified of the channel as
Hc�f �=Hf �f �d�Ht�f ��t� ·Hr�f ��r� (4)
where Hs�f ��s� (s = n or m) represents a complex of channel
between transmitted antenna and received antenna with differ-
ence polarization the �s = ��s��s� as
Hs�f ��s� = Hs�f ��s��s�
= �sHs�f ��s��s + �sHs��f ��s��s� (5)
which has the relation as
1
4�
∫ 2�
0
∫ �
0
�Hs�f ��s��s��2 sin� d� d�= �s (6)
where �s denotes the antenna efficiency, in addition, the solution
can be normalized with isotropic antenna.
3.2. UWB Correlation Receiver
The UWB correlation receiver is considered by using template
waveform, which is shown in Figure 3. The template received
signal and isotropic received signal are analyzed based on input
and output conrrelation waveforms are considered. In this paper,
Fig. 2. The experimental setup.
3290
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3289–3292, 2015
Fig. 3. System modeling of UWB system.7
the optimal receiver is proposed from optimum signal with
received signal of UWB channel.
In this section the condition of noise channel between are
discussed, the transmission channel of receiving signals and
isotropic signal templates, Ho�f � and Hi�f � are prospectively
normalized as∫ �
−��Ho�f ��2 df =
∫ �
−��Hi�f ��2 df = 2fb (7)
In addition, the condition of noise bandwidth is formulated
by N0fb , where N0/2 represented additive white Gaussian noise
(AWGN).
The UWB optimum receiver with received signal and isotropic
signal templates is written as
Hopt�f �=√
2fbV∗r �f �√∫�
−� �Vr�f ��2 df(8)
Hi�f �=√
2fbV∗r−i�f �√∫�
−� �Vr−i�f ��2 df(9)
where �·�∗ the free space complex conjugate, Vr�f � and Vr−i�f �denote the spectrum frequency of receiving signals of channel
measured and isotropic antennas respectively. Then, the expres-
sion from can be shown as
Vr�f �=Hc�f �Vt�f � (10)
Vr−i�f �=Hf �f �Vt�f � (11)
where Vt�f � represents the spectrum waveform as related with
Fourier transform
Vt�f �=∫ �
−�vt�t�e
−j2�f t dt (12)
Hc�f � represents the channel vector as from Section 3.1 and
Hf �f � is, which given by
Hf �f �d� =c
4�d�f � e−j2�fd/c (13)
where d denotes distance of between transmit and receive
antenna and c is the velocity of light.
3.3. Transmission Gains
The UWB signal distortion to evaluate the peak value of the cor-
relation receiver output of biconical antennas simplified that the
received signal and isotropic signal templates are compared. The
waveform distortion from antenna is normalized with correlation
template receiver, the UWB transmission gain is represented in
the UWB transmission gain of the signal-to-noise ratio (SNR) at
the UWB impulse radio.
Fig. 4. UWB measurement gain of rectangular by using FCC underbiconical–biconical antennas transmission link.
Therefore, UWBIR transmission of correlation receiver as tem-
plate waveform GWM is written by
GWM =max
∣∣∣∣∫�−� vr �t�hWM�t−��dt
∣∣∣∣max
∣∣∣∣∫�−� vr−iso�t�hWC�t−��dt
∣∣∣∣(14)
The gain of UWBIR transmission and the isotropic template,
where GWC as
GWC =max
∣∣∣∣∫�−� vr �t�hWC�t−��dt
∣∣∣∣max
∣∣∣∣∫�−� vr−iso�t�hWC�t−��dt
∣∣∣∣(15)
Fig. 5. UWB measurement gain of RRC by using indoor FCC underbiconical–biconical antennas transmission link.
3291
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3289–3292, 2015
Fig. 6. UWB measurement gain of RRC by using outdoor FCC underbiconical–biconical antenna transmission link.
4. RESULTS AND DISCUSSIONThe evaluation of waveform distortion in the UWBIR transmis-
sion model are considered of transmission gain are shown. First,
the UWB transmission gain of receiving a signal of receiving
signals and isotropic signal templates at the receiver are consid-
ered. The UWB measurement gain of two passband waveform is
shown. In particular of FCC, the rectangular passband, the RRC
passband verifying indoor scenario and outdoor surrounding are
in Figures 4 to 6, respectively. The comparison of experimen-
tal results is compared to receive signals and isotropic templates
following FCC, which can know clearly that using a common
frequency band is better than FCC. At the FCC, the estimated
resultant of the rectangular passband, the RRC passband com-
pare both indoor scenario and outdoor surrounding as 1.21, 1.32
and 1.27 dB as completely.
5. CONCLUSIONIn this paper, the biconical antennas are investigated as a broad-
band antenna to educate for UWB communication system, also
waveform distortion due to channels and antenna measurement
has been proposed. The waveform distortion of channel transfer
function and biconical antennas are evaluated for UWB impulse
radio system by using extended Friis transmission formula. The
correlation receiver with received signal and isotropic signal tem-
plates are evaluated. Using a biconical antenna is applied as the
transmit and receive antennas. As the results, the relative gains in
the received signal and isotropic signal templates are every small
difference.
References and Notes1. G. Adamiuk, UWB antenna for communication systems, Proceeding of the
IEEE, April (2012), Vol. 100, pp. 2308–2321.2. V. Yajnanarayana, Design of impulse radio UWB transmitter for short
range communication using PPM signals, IEEE International Conference onElectronics, Computing and Communication Technologies, January (2013),pp. 1–4.
3. J. Farserotu, A. Hutter, F. Platbrood, J. Gerrits, and A. Pollini, Wireless Per-sonal Communications 197 (2002).
4. H. T. Friis, A note on a simple transmission formula, Proc. IRE, May (1946),Vol. 34, pp. 254–256.
5. J. Takada, S. Promwong, and W. Hachitani, Extension of Friis’ transmissionformula for UWB systems, Technical Report of IEICE, WBS2003-8/MW2003-20, May (2003).
6. S. Promwong, W. Hachitani, and J. Takada, Experimental evaloation schemeof UWB antenna performance, Technical Meeting on Instrument and Mea-surement, IEE Japan, IM-03-35, June (2003).
7. S. Promwong, W. Hachitani, J. Takada, P. Supanakoon, and P. Tangtisanon,Experimental study of ultra-wideband transmission based on Friis’ transmis-sion formula, The Third International Symposium on Communications andInformation Technology (ISCIT) 2003, September (2003), Vol. 1, pp. 467–470.
8. S. Promwong, J. Takada, P. Supanakoon, and P. Tangtisanon, Path lossand matched filter gain for UWB system, 2004 International Symposium onAntenna and Propagation (ISAP), August (2004), pp. 97–100.
9. S. Promwong, J. Takada, P. Supanakoon, and P. Tangtisanon, Path loss andmatched filter gain of free space and ground reflection channels for UWB radiosystems, IEEE TENCON 2004 on Analog and Digital Techniques in ElectricalEngineering, November (2004), pp. 125–128.
10. F. Tufvesson and A. F. Molisch, Ultra-wideband communication using hybridmatched filter correlation receivers, 2004 IEEE 59th Vehicular TechnologyConference (VTC), May (2004), Vol. 3, pp. 1290–1294.
11. P. Supanakoon, K. Teplee, S. Promwong, S. Keawmechai, and J. Takada,Theoretical SNR gain and BER performances of UWB communications withmatched filter and correlation receivers, The International Technical Con-ference on Circuits/Systems, Computers and Communications (ITC-CSCC2006), July (2006) pp. 269–272
12. P. Supanakoon, K. Wansiang, S. Promwong, and J. Takada, Simple waveformfor UWB communication, The 2005 Electrical Engineering/Electronics, Com-puter, Telecommunication, and Information Technology International Confer-ence (ECTI-CON 2005), May (2005), pp. 626–629.
13. Federal Communications Commission, Revision of Part 15 of the Commis-sion’s Rules Regarding UWB Transmission Systems, First Report, FCC 02-48,April (2002).
14. W. Hirt and M. Weisenhorn, Overview and implications of the emerging globalUWB radio regulatory framework, Proceeding the 2006 IEEE InternationalConferences on Ultra-Wideband, September (2006), pp. 581–586.
15. J. Foester, Channel Modeling Sub-Committee Report Final, IEEEP802.15-02/368r5-SG3a, November (2002).
Received: 13 October 2014. Accepted: 27 November 2014.
Sonny Zulhuda1�∗, Ida Madieha Abdul Ghani Azmi1, and Nashrul Hakiem2
1Civil Law Department, International Islamic University Malaysia2Faculty of Science and Technology, UIN Syarif Hidayatullah Jakarta
Many hypes are currently surrounding the “datafication” such as the Big Data, Cloud and BYOD. The proliferationof data from ubiquitous sources is often not counter-balanced with adequate awareness and prudent risk man-agement by the end-users, making it easier for others to take advantage of the new technology and reap fromthe abundant data available for all kinds of purposes including criminal. IT stakeholders should view Big Datanot only as a new exciting technological advancement, but also a frontier full of potential risks to be addressednot only by industrial best practices, but also by the reforming laws in the area of information security and dataprivacy. This paper sets to undertake two major tasks. Firstly, examining the types of legal risks involved in theBig Data environment. Secondly, it highlights some aspects of data privacy and security laws already containedin the current data protection laws in Malaysia.
Keywords: Datafication, Big Data, Data Security, Personal Data Protection, Malaysia.
1. INTRODUCTION“Datafication” is, in the words of Meyer-Schonberger and Cukier,
the process of quantifying all information around us: our loca-
tion, movement, communications, usage of devices, etc. which
will allow us to use such information in new ways, such as in pre-
dictive analysis. This will help us further to unlock the implicit,
latent value of the information.1 The amount of those quanti-
fied data may prove to be unimaginable, hence the expression
“Big Data.”
“Big Data” is an expression that typically refers not only to
specific, large datasets, but also to data collections that consoli-
date many datasets from multiple sources, and even to the tech-
niques used to manage and analyze the data.2 According to the
report by a US official report, Big Data is big in two differ-
ent senses. It is big in the quantity and variety of data that are
available to be processed. And, it is big in the scale of analysis
(“analytics”) that can be applied to those data, ultimately to make
inferences.3
It is an ocean of data out there where we can only swim or
sink.4 This datafication is poised to reshape the way we live,
work and think.1 Those data are abundant. It comes in a mas-
sive volume, velocity and variety that are unprecedented.5 In the
words used by the Gartner IT glossary, Big Data is defined as
“high-volume, high-velocity and high-variety information assets
∗Author to whom correspondence should be addressed.
that demand cost-effective, innovative forms of information pro-
cessing for enhanced insight and decision making.”6
Another immediate effect of datafication is the emergence of
“cloud computing.” The Big Data environment has in turn led
the companies to tap server capacity as needed to accommodate
an enormous scale required to process big datasets and run com-
plicated mathematical models. Here is the nexus between Big
Data and cloud computing.7 The massive dataset that an organ-
isation is having at its disposal presents another challenge: how
can it store, process and exploit such big data in the best and
most efficient manner? This is what makes cloud computing an
ultimate choice. Being hailed as the future of information tech-
nology (IT) architecture, by the year 2018 cloud computing is
projected to be a major medium for delivery of information and
other IT functions at both the consumer and corporate ranks.8
The US National Institute of Standards and Technology (NIST)
defines cloud computing as “a model for enabling ubiquitous,
convenient, on-demand network access to a shared pool of con-
figurable computing resources. That can be rapidly provisioned
and released with minimal management effort or service provider
interaction.”9 Hoover reckoned that the businesses increasingly
see cloud computing as a “valuable proposition for decreas-
ing technology costs, enabling and accelerating the delivery of
new technology services, and refocusing technology workers on
mission-oriented tasks that deliver more business value than time
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3347–3351, 2015
Nevertheless, cloud computing is not a risk-free concept as
it also has a tremendous implication on what sets of new risks
an organisation will encounter in their data management. This
is because the main issue with a cloud computing is that the
owner or controller of the data has no longer assumed a complete
and exclusive control over their data assets: some portions of
the analytics and processing are being outsourced to an external
entity.
Meanwhile, this Big Data environment compels companies to
be become increasingly agile, productive and focused in-order
to achieve a competitive advantage in the market place. This
development pushes for renewed working options such as resort-
ing to cloud computing and IT service outsourcing. Besides that,
to create more fluidity and flexibility at work, employers allow
staffs to work virtually from any place but the office itself. They
typically tell the employers that they can bring their own work
device to and from home—hence the “Bring Your Own Device”
(“BYOD”) concept.
This paper is making an attempt to revisit this aspect, namely
the risks of datafication to an organisation, more specifically on
the practices of cloud computing and BYOD and to analyse it
from the legal perspective. It employes a legal doctrinal research
that collects primary data from the statutory/legislative provi-
sions and secondary data from case reports. It observes cases
where data resources management was found problematic, then
explores the relevant legal and regulatory requirements on the
area of personal data protection retrieved from the Malaysian
legal framework.
The contribution brought about by this paper is in its strategic
objective: it seeks to prove how the big data environment today
gives rise to intersection with data protection legal requirements.
It highlights the need for a future research on organisational
road-map to comply with data protection requirements in more
structured and systematic manner as part of organisation’s data
governance.
2. CHALLENGES OF PERSONAL
DATA MISMANAGEMENTGiven the promises of datafication, companies and individuals
are quickly trying to grab every possible fortune that pops up
in this process. However the process of embracing the Big Data
environment is often marked with its dark side. The individu-
als whose personal information have increasingly flooded public
spaces and open network are normally the immediate receiver of
the consequences. Their data are thus exploited: discreetly stored,
aggregated, disclosed, exchanged for money, traded at black mar-
ket or kept indefinitely. Thus it is natural that each stakeholders
of this Big Data are expected to manage their data well so as to
avoid those risks.11
The dark side to Big Data takes many forms, such as an
interference with privacy.12 For example, the extensive amounts
of personal information revealed during online transaction have
taken the relationship between customer profiling, predicting
trends, and marketing to a whole different level. As Reyhaneh
explains, Big Data is capable of tracking movements, behaviour
and preferences, and predicting the behaviour of individuals with
unprecedented accuracy. The more access business has to Big
Data, the better it can target us with advertising that matches
(or predicts) our specific interests. This is, however, often done
without our consent.12
As for the cloud computing, risk appears as early as it starts.
The transfers of organisations’ databases to external and central-
ized data centers means the transfer of certain duty of safeguard-
ing the data itself. Doubt would naturally grow when the security
measures are taken over by cloud providers, the latter may not
be fully trustworthy.8 Thus, with the explosion of data being
outsourced to this external parties, the increasing prevalence of
identity theft and data security breaches for cloud consumers is
paramount.8
Understanding the legal requirement on data security and pri-
vacy would be easier if one knows the evils and mischief of
data mismanagement. In the perspective of information gover-
nance and information security, personal data is an asset because
it has value. Outsiders would be interested to have access to its
customer database including external marketers, business com-
petitors or any adventurous individuals who grab any opportunity
to make use of such company’s database.
One can borrow cases from the UK’s Information Commis-
sioner Office to show data user’s mishandling of their personal
data resources in the following paragraphs.13
• It was reported in 2008 that the UK Financial Services Author-
ity (FSA) found three units under HSBC group had failed to
put in place adequate systems and controls to protect customers’
details from being lost or stolen. HSBC Life, HSBC Actuaries
and HSBC Insurance Brokers were fined a collective amount of
over 3 million pounds for having lost unencrypted disks contain-
ing personal details of their customers.
• In 2011, an encrypted memory stick containing patients’
sensitive personal data of the Arthur House Dental Care was
accidentally lost from the possession of one of the dentist. The
memory stick was used as a temporary back-up solution and was
taken home by a dentist for safekeeping. The memory stick was
found in public and was sent to the Information Commissioner
Office (ICO), who in turn took action against the dental care for
their failure to place proper safeguards in providing data backup
system.
• An e-mail containing a medical report of patient’s health was
mistakenly sent by Aneurin Bevan Health Board (ABHB) to a
wrong patient in Wales. The UK ICO found that there was not
enough robust system to prevent this case of accidental disclo-
sure. As part of enforcement notice, the data user had to imple-
ment some measures including staff data protection training,
monitoring of compliance, and the introduction of new checking
processes before personal information is sent out.
• A contractor company who was tasked by Brighton and Sus-
sex University Hospital to remove and destroy old computer
hard drives containing sensitive personal data was found to be
in breach of Data Protection Act after one individual from the
contractor company took some of the hard drives and sold them
on eBay. The ICO has levied a fine of £325�000 on BSUH over
the breach for failure to put proper supervision on the process of
data removal and on the involvement of a third party contractor.
The cases shown above, dealt with under the UK Data Pro-
tection Act 1998 should serve as an alarm to all data users. The
emerging legal rule under the new Act places a higher standard of
due diligence to be complied by companies. Previously, a com-
pany whose data was found to have unintentionally been leaked,
lost, or subjected to disclosure would only be potentially liable
if there is an evidence of negligence in their part, which likely
means establishing positive breach of duty of care. With new
3348
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3347–3351, 2015
written laws emerging in many jurisdictions, such standards had
become more certain and measurable, some of which is portrayed
in this paper.
3. DATA PROTECTION LAW IN MALAYSIAMajor legal issues on data privacy in Malaysia were addressed
by the Personal Data Protection Act (PDPA) 2010. Being the
main legal framework for protecting data privacy of individuals,
PDPA regulates the processing of personal data in commercial
transactions.14 Under Section 4, “personal data” refers to any
“data that relates directly or indirectly to a data subject, who
is identified or identifiable from that information or from that
and other information in the possession of a data user, including
any sensitive personal data and expression of opinion about the
data subject.” Meanwhile, “commercial transactions” mean “any
transaction of a commercial nature, whether contractual or not,
which includes any matters relating to the supply or exchange of
goods or services, agency, investments, financing, banking and
insurance.”
The enactment of the PDPA is arguably a milestone for the
development of e-commerce in Malaysia, considering that a mas-
sive and increasingly valuable amount of personal information
are being stored, processed and exploited.14 At the heart of PDPA
is a set of duties under the data protection principles from which
stemming all the rights, duties and liabilities of each of data user
and data subject (“data user” is those who use, collect, process,
etc. the personal data that belong to certain individuals, i.e., the
“data subject”). There are generally seven categories of the duty
spelled out in Table I, whereas each of those principles con-
tributes to the data protection in Malaysia.
The most relevant provision under the Act would be the
offence of unlawful collecting or disclosing of personal data
(section 130). If any person is found to have knowingly or reck-
lessly collected or disclosed personal data that is held by the
data user without the consent of the latter commits an offence
punishable with a fine of maximum MYR500, 000.00 or with
imprisonment for a maximum term of three years or with both.
The same penalties await those who sell personal data under the
circumstances set out above. This provision does not specify the
manner of such collection, disclosure or selling of the personal
data but instead leaves it open so as to be able to catch offenders
Table I. PDP principles under the PDPA 2010.
Section Principle Information security implications
6 General principle No process of personal data which isexcessive and/or without the consent ofdata subject.
7 Notice and choice Proper notification on the purpose of thatdata collection/processing
8 Disclosure Prohibits unauthorized disclosure orsharing of personal data
9 Security Imposes security measures by data usersthat commensurate the risk of securitybreach
10 Retention Personal data shall not be keptunnecessarily
11 Data integrity Right of data subjects to correct andupdate their personal data
12 Data access Right of data subject to have an access tohis own personal data the at the user’sdatabase
in various ways or modus operandi. The provision, it is argued is
useful in providing strong and effective measures against unlaw-
ful collection and disclosure of personal data.
Another important provision is the duty of data users such as
those cloud service providers to conduct due diligence as to the
reliability and security of their electronic system. This is because
under section 133, the board of director or any officer responsible
for the management of company may be charged for an offence
by body corporate, unless he can prove his absence of knowledge,
and that he had taken all reasonable precautions and exercised
due diligence to prevent the commission of the offence. Given
this analysis, it can be said that the PDPA can lend a hand for
the security of data in the electronic environment.
4. CLOUD AND DATA PROTECTIONCloud computing can be understood as a way of delivering
computing resources as a utility service via network, typically
the Internet, scalable up and down to user requirements.15 The
major feature of cloud computing is that it enables decentral-
ization of resources such as data and applications system, thus
would enable an entity (e.g., companies) to free their resources
or at least be more focused on the functions and less on man-
aging their data resources. As a result, the data owner’s control
over their own data would substantially decrease while surrender-
ing more control to a third party to manage their data resources.
As such, a new risk is just getting created, namely the increasing
risk of data security and confidentiality breach.15
From data protection perspective, the core challenge in cloud
computing and IT service out sourcing would be the fact that data
user loses a complete control over personal data of their data sub-
jects to whom they are initially answerable. Huge amount of per-
sonal data will be processed by a third party on their behalf. With
the absence of immediate supervision and the physical bound-
aries, how can data users ensure the security of personal data?
In Malaysia, the legislature has stepped forward to put cer-
tain legal requirements that will change the landscape of cloud
computing services.14 Under the PDPA, those third parties who
provide cloud computing or other IT services are called “data
processor” simply because they process the personal data solely
on behalf of the data user, and do not process the data for any
of their own purposes. The moment they process it for their own
purpose, they become a “data user” for that particular purpose.
The following is a set of duties and responsibilities of a data
processor that can be derived or extracted from the provisions of
the PDP Act 2010:
(a) The security principle in section 9 mandates that data user
shall take measures for ensuring the reliability, integrity and
competence of personnel having access to the personal data.
This in turn requires the establishment of an agreed and well-
defined roles and responsibilities between data processor and
data user in relation to the processing of personal data. This can
be achieved with a well-crafted PDPA-compliant service level
agreement (SLA);
(b) Following the above based on the same provision, it is there-
fore desirable that data users and data processors have in place
internal policies, standards or procedures of information security
measures that govern the processing of data (including that of
protective, detective and responsive measures). These measures
include technical and organizational ones.
3349
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3347–3351, 2015
(c) Based on section 133 on the offence by body corporate that
has been mentioned earlier, the data users and data processors
will have to implement due diligence to ensure and maintain
compliance with the contractual guarantees and internal mea-
sures above. This “due diligence” may be generally interpreted as
putting in place a proper system of data management and proper
monitoring of such system (See: Universal Telecasters (Qld) Ltd.v Guthrie [1978] FCA 18);
(d) By virtue of section 101 of the PDPA 2010, data users and
data processors are both subject to inspection and compliance
audit by Commissioner to the information system used for the
processing of personal data.
(e) On top of that, there are more stringent rules when it involves
a foreign party as cloud provider. The PDPA requires a two-
layer safeguard including authority clearance and contractual
safeguards. This is provided in section 129 of the PDPA 2010.
Based on the above, it is important to reiterate here that in the
first place it is the duty of every data user to ensure that such
PDPA-compliant agreement takes place between them and data
processors. The obligations of data processors such as the cloud
service providers will necessarily reflect those duties of the data
users, as long as the cloud providers process the data on behalf
of the data users, and not for their own purpose.
5. BYOD–“BRING YOUR OWN DEVICE”Another recurring challenge with personal data management is
about the trend of “bringing your own device” or BYOD. It is
a commendable development that employees are more mobile
and even free to bring their own device to work from home or
anywhere else but their office. However, BYOD can be risky if
one does not execute proper threat analysis and risk management
measures. The above cases of personal data breaches due to the
lost or stolen devices should serve as lessons for all of us.
Under the security principle, the PDPA clearly mandates that
data user shall, when processing personal data, take practical
steps to protect the personal data from any loss, misuse, modifi-
cation, unauthorized or accidental access or disclosure, alteration
or destruction. By virtue of section 9 of PDPA 2010, such prac-
tical steps should take into account:
• The nature of the personal data and the harm that would result
from such loss, misuse, modification, unauthorized or accidental
access or disclosure, alteration or destruction;
• The location where the personal data is stored;
• Any security measures incorporated into any equipment in
which the personal data is stored;
• The measures taken for ensuring the reliability, integrity and
competence of personnel having access to the personal data; and
• The measures taken for ensuring the secure transfer of the
personal data.
Given the above, the authors outline the following as necessary
safeguards to ensure BYOD works well for the organisations in
big data environment:
(a) Identity data asset classification. This means data users
should be aware of the nature of the harm on each of class of
personal data in their possession. Only then they can determine
which personal data should be kept at which security level.
(b) Ensure and manage proper access control. In other words,
data users should identify who can bring home what device and
what data. Data users should therefore apply the principle of
least privileged.
(c) Distribution of role between employees. This is important
because not all employees need to bring their devices home.
(d) Risk management, business continuity plan and disaster
recovery management. This is because the Act requires data users
to carefully determine the right measures taken for ensuring the
secure transfer of the personal data.
(e) Due care and diligence: On top of what has been mentioned
on similar issue in the earlier passage, it is noteworthy that the
above measures are collective responsibility, not only that of the
IT division in an organization. The precaution and due diligence
to be taken by all members in an organisation is to encourages
a massive overhaul in redefining each role and responsibility
required in the data processing activities.
6. CONCLUSIONGiven the analysis above, we can conclude that the data protec-
tion law has been developed in major parts to respond to the chal-
lenges of datafication and trends associated with it. In Malaysia,
this attempt is arguably spearheaded by the Personal Data Pro-
tection Commissioner established under the PDP Act 2010. The
personal data law mandates certain steps of legal risks man-
agement to be taken by business entities in securing their data
resources. Compliance with the law helps organizations in safe-
guarding their assets in this Big Data environment.
Personal data law is not only concerned with protecting “per-
sonal” interest such as the privacy of individuals, but it is also
about safeguarding organizational assets in the whole lifecycle
of such data. Such law oversees not only on the lifecycle of
organizational data asset, but also the whole processes of IT
governance involving people, process and technology. Further-
more, it is not only about data management, but also personnel
awareness, robust system protection, and good business and gov-
ernance (such as the requirement of data due diligence). Com-
pliance with data privacy law is arguably a passport to a global
business because personal data protection is indeed becoming
another widely-accepted indicator of transparency and good gov-
ernance and can be potential trade barrier in the future.
This paper reiterates that today’s business landscape has
changed. With data increasingly becomes bigger and faster, risks
and threats need to be more comprehensively managed. Technol-
ogy alone is not an answer. It has to be further accompanied by a
paradigm shift in the mindset of people and processes involved.
The more vigilant one is, the more he can control the risks and
challenges posed by the big data environment.
It is reiterated here that this paper undertakes the task to estab-
lish the link between requirements of protecting personal data
as mandated by the data protection law with the contemporary
trends in big data environment today. It also paves the way for-
ward to highlight the need for a future research on organisational
road-map to comply with data protection requirements in more
structured and systematic manner as part of organisation’s data
governance.
References and Notes1. V. Mayer-Schonberger and K. Cukier, Big Data: A Revolution That Will Trans-
form How We Live, Work and Think, John Murray, London (2013).2. M. R. Wigan and R. Clarke, IEEE Computer 46 (2013).
3350
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3347–3351, 2015
3. Big Data and Privacy: A Technological Perspective, President’s Council ofAdvisors on Science and Technology. May 2014. Available at: https://www.whitehouse.gov/sites/default/files/microsites/ostp/PCAST/pcast_big_data_and_privacy_-_may_2014.pdf. Accessed April 30, (2015).
4. Intel, What Happens in an Internet Minute?, 2014, Available at: http://www.intel.com/content/www/us/en/communications/internet-minute-infographic.html. Accessed April 30 (2015).
5. W. H. Davidow, Overconnected: The Promise and Threat of the Internet, Del-phinium Books, New York (2012).
6. ICO. Big data and data protection. 20140728. Version: 1.0. Available at:https://ico.org.uk/media/for-organisations/documents/1541/big-data-and-data-protection.pdf. Accessed April 30 (2015).
7. HM GovernmentHorizon Scanning Programme, 2014, Emerging Tech-nologies: Big Data, TheEmerging Technologies Big Data Community ofInterest. December 2014. Available at: https://www.gov.uk/government/
uploads/system/uploads/attachment_data/file/389095/Horizon_Scanning_-_Emerging_Technologies_Big_Data_report_1.pdf. Accessed April 30 (2015).
8. J. A. Harshbarger, Journal of Technology Law and Policy (J. Tech. L. andPol’y) 16, 229 (2011).
9. J. Ryan, Santa Clara Law Review 54, 497 (2014)
10. J. N. Hoover, Journal of Business and Technology Law (J. Bus. and Tech. L.)8, 255 (2012).
11. N. Gifford, Information Security–Managing the Legal Risks, CCH AustraliaLtd., Sydney (2009).
12. R. Saadati and A. Christie Internet Law Bulletin 67 (2013).13. ICO. UK’s Information Commissioner’s Office. 2015. Available at: https://
ico.org.uk/. Accessed April 30 (2015).14. A. B. Munir and S. H. M. Yasin, Personal Data Protection in Malaysia: Law
and Practice, Sweet and Maxwell Asia, Selangor (2010).15. C. Millard, Cloud Computing Law, Oxford University Press, Oxford (2013).
Received: 23 November 2014. Accepted: 11 January 2015.
Muhammad Suhaizan Sulong1�∗, Azlianor Abdul-Aziz1, Andy Koronios2, and Jing Gao2
1Faculty of Information and Communication Technology, Universiti Teknikal Malaysia Melaka, 76100, Malaysia2School of Information Technology and Mathematical Sciences, University of South Australia, Australia
Enterprises are now seeking to cater the needs of their business change to improve business agility and toincrease information technology flexibility. They have transformed with the initiative of implementing serviceorientation from their existing information technology infrastructure to a managed services environment. Thisrequires a step-by-step process in the implementation of the service orientation initiative. In this paper, the four-phase process of implementing service orientation initiative derived from the related literature is taken to thecase study of seven case enterprises to check for its applicability in wider enterprise settings. It has resultedthat this four-phase service orientation initiative process has been clearly employed and agreed upon by mostof the case enterprises. It is therefore believed that the four-phase process can be a generic initiative processof implementing service orientation for any enterprises to maintain viability in a services environment.
Keywords: Service Orientation, Enterprise Services, Service-Oriented Architecture.
1. BACKGROUND STUDYMany enterprises nowadays have deployed and integrated multi-
ple information systems into their core business operations across
different departments.1 These information systems frequently
shared and produced similar information to support business
objectives across the departments. This shows that these informa-
tion systems implemented are impractical and developing more
of this can lead to high costs. Besides, as business conditions
and environment changes with evolving consumer needs, the
enterprises are seeking competitive advantages through advanced
information systems that should reflect the changes.2 The current
information technology (IT) architecture that is siloed mololithic
in style could not support the advanced IT requirements.3 Thus,
rapid changes in the business environment have made planning
for new architecture more important and are required to reduce
unnecessary costs from using the existing information systems.
The new architecture is referred to a service-oriented architecture
i.e., service orientation in which enterprises that are having an
initiative to implement this would increase IT flexibility as well
as improve business agility.3
Service orientation as well as service-oriented architec-
ture (SOA) has been a buzzword for enterprise seeking services
environment. A clear understanding of what service orientation
means is important from two perspectives. From a technical
perspective, it can be referred to as a technology for enabling
∗Author to whom correspondence should be addressed.
business4 which is defined as “architectural style where systems
consist of service users and service providers.”5 While from the
business perspective, it can be referred to as aligning technol-
ogy with business,6 which is defined as a paradigm or a way
of thinking or designing information systems.7 We consider both
definitions for this study.
Implementing service orientation initiative at the enterprise
level needs to be in phases through a step-by-step approach8
which means this enterprise initiative transforms to services envi-
ronment that can quickly respond to business change. A recent
study has reviewed the related literature on the service orientation
initiative extensively.9 From the study, five service orientation
initiatives are discussed and evaluated (refer to Table I). Both
Oracle10 and IBM11 have, in reality, been in the service orienta-
tion arena for a long time and provide own unique approaches
to comprehensive service orientation in which they retrofit SOA
into their established computing platforms. Gartner Research
also extends its specific approach namely Application Activity
Cycle for SOA12 to implementing service orientation initiative.
Although the service orientation marks an active area of devel-
opment in industry, the researchers in academia also intensified
their efforts to realise the potential of service orientation in which
they proposed SOA Roadmapping13 and SOA Adoption Man-
agement Roadmap.14 It is worth noting that in order to increase
confidence of the review result, a greater variety of service ori-
entation initiatives is covered.
After having carefully reviewed and compared these five
service orientation initiatives, they have a high integrity and
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3356–3359, 2015
consistency regarding the service orientation process and thus,
the general four phases9 are derived:
• Strategic Planning: Starting point of an enterprise to plan
strategically its service orientation initiative.
• Building Architecture: Designing and developing services and
infrastructure to integrate with existing operating platforms.
• Services Operation: All developed services are being exposed
to allow integration with other systems and services.
• Continuous Improvement: Service orientation is at the point
of improvement either changing current services or creating new
services to accommodate business change and measure its per-
formance.
The name of the second phase is suggested to rephrase to
“Architecture Building” for a proper terminology.
2. RESEARCH METHODOLOGYThis study used a multiple case study design15�16 that involve
analysing real-world case situations of seven enterprises with var-
ious types of business in order to seek how these enterprises
implementing SOA initiative to increase their business agility.
From these seven enterprises, 28 participants, in their respec-
tive SOA team members grouped into three levels; management,
architectural and delivery, were gave written informed consent
for the semi-structured interviews to gain in-depth insights into
how they experienced the service orientation initiative process.
These seven enterprises were selected on an opportunistic basis
which means the selection based on any invited enterprises that
having a service orientation initiative and agreed to participate in
this study.
The format of each interview was one to one session lasted
approximately an hour, which sought to create a relationship
between the researcher and the participant, providing dynamism
Table I. Comparison of service orientation initiative process.9
Service orientationStudy initiative process
Oracle’s approach to SOA10 1. Establish a strategic plan for SOAadoption; 2. execute the SOA programlevel activities; 3. deliver projects andservices following SOA best practices;4. establish ongoing guidance andgovernance.
IBM scope of SOA adoption11 1. Ad hoc stage; 2. technology adoptionstage; 3. line-of-business adoptionstage; 4. enterprise adoption stage;5. value-net adoption stage
Gartner application activitycycle for SOA12
1. Strategise: SOA adoptionconsiderations; 2. evaluate: planningand designing SOA systems;3. execute: implementing andmanaging SOA in the real world;4. review: improving and refining theuse of SOA; 5. innovate
SOA roadmapping13 1. Planning and analysis; 2. design andconstruction; 3. deployment andoperations; 4. management andgovernance
SOA adoption managementRoadmap14
1. Set strategy organization; 2. planningstrategy; 3. organization and business;4. operations planning; 5. design;6. implementation; 7. monitoring andtesting; 8. establishment and feedback.
and flexibility in the discussion.17 Table II presents the back-
ground of participating enterprises, including their service orien-
tation initiatives (implementation year, platform and system) and
volunteered participants (job title).
The analysis was assisted with the use of analytical software
tool called NVivo.18 This tool can assist in categorising and
analysing the interview data upon transcription. The analysis pro-
cess includes first is to transcribe the interview data. Next, using a
thematic approach, the interview data can be categorised and then
interpreted towards the research objective. The findings across
the cases regarding the process of their service orientation initia-
tive implementation are presented in the next section.
3. RESEARCH FINDINGS DISCUSSIONThe general four-phase of the service orientation initiative
E Initiated 2010 on Microsoftplatform; A federally fundedproject for implementingSOA specific in GISsystems for naturalresource data.
Chief Information Officer;Operations Manager;GIS Administrator
F Initiated 2008 on software AGand IBM platform; To linktogether disparatesystems—To support theacquisition strategy forfuture integrationrequirements.
Clearly statedthat—The fourphases tookplace toimplement all ofthe top-downservices thatwere required.
Use the IBM’smethod—Service-OrientedModelling andArchitecture(SOMA).
Via an outsourcedcompany usingthe Microsofttechnology.
initiative was agreed upon by most cases. Although the findings
can be depends on the implementation platform, most respon-
dents stated that their service orientation initiatives have imple-
mented through this general approach.
That makes a great deal of sense. We fell into a holebetween planning and building. We have plan, buildand operate it all happening at once and continu-ous improvement too. � � �There are still many otherthings that we are, I guess planning and building,and of course the operational team is still working.So there’s actually the operations guys are contin-ually improving their capabilities as well.
Enterprise Infrastructure Architect (Case A)
The phases, yeah, we have done some planning ina very limited area of the integration. We have builtsome services on the hub so that they can be usedand we have actually deployed them and we areusing them. Have we continuously improved them?Not to my knowledge.
Team Lead Solution Design (Case C)
Well, we went through these different phases andwe experiment the four phases, because we couldonly trying to extend the model to generalise all theservices of all applications.
Integration Program Manager (Case D)
We’d do that. So you will do your plan. Youwill then determine, understand what services, etc.you’re going to execute on. You will then create theservice, implement, measure your ROI.
Senior Development Manager (Case F)
Yeah, it pretty much happened like that. Like theprobably the only thing for a particular system ora particular service, [we] tried to implement all ofthe top-down services that were required.
Application Architect (Case G).
Only two cases—Case B and E, indicated they had used SOA
vendor’s methodology to implement their service orientation ini-
tiative. For the case B enterprise, as they used IBM for their
service orientation initiative implementation, the methodology
that IBM provides was SOMA—Service-Oriented Modelling and
Architecture. According to the respondent:
We’re definitely looking to continue to implementSOA in terms of our progress along the SOA model,
we have done a lot of things to try and mature theway that we use SOA. So we use the SOA method-ology, SOMA. � � � If you follow �our� lifecycle for aproject you get an initiation, where people are try-ing to think of a technology solution. They then gointo a defined phase, where they’re trying to findwhat that solution is, and then they go into an anal-ysis and design phase. So we’re a delivery team—allwe do is to deliver.
Head of Integration Services (Case B)
Besides, in the Case E enterprise, their service orientation
initiative has implemented using Microsoft technology with the
assistance of an outsourced company as the respondents stated:
This organisation was ripe for an SOA, for animplementation of an SOA to meet some departmen-tal objectives and government objectives and so thatwas really what initiated that. It was actually out-sourced to an Indian company. So with regards tothe actual SOA technology and infrastructure that’sbeen built on, I was making sure when they weredeveloping and designing its functionality, it metour needs.
GIS Administrator (Case E)
It’s trying to be Microsoft based at this stage, whichis Microsoft Dynamics.
Operations Manager (Case E)
Although the enterprises of Case B and E followed IBM and
Microsoft’s approaches in implementing their service orientation
initiative respectively, the approaches also have the generic pro-
cesses like capturing user requirements during the product design.
Table III shows the summary of cases that clearly stated in
following the general approach and cases that have different
approaches in implementing their service orientation initiatives.
Therefore, it is clear that the proposed general approach via the
four-phase implementation can be employed for implementing
service orientation initiatives. One important implication is that
any enterprises who are still pursuing their service orientation
initiatives can consider the use of this general approach through-
out its implementation process in order to become SOA-enabled
enterprises.
4. CONCLUSIONService orientation is significant to enterprises in order to
improve business agility and IT flexibility as well as achieve the
3358
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3356–3359, 2015
business value of SOA. The service orientation initiative pro-
cess is developed generally by comparing five different service
orientation initiatives studied previously. The general phases are
strategic planning, architecture building, services operation and
continuous improvement. From the Table III, five out of seven
case enterprises have affirmed that they applied the general ser-
vice orientation initiative process when implementing their ser-
vice orientation initiatives. Detail explanation for each phase
according to the case study is as follows:
• Strategic Planning: This first phase is where an enterprise
starts to plan its business transformation initiative to service ori-
entation. It involves top-management level that is significant for
supporting the entire initiative implementation in order to achieve
specific business needs. With an implementation strategy that fits
well within the enterprise settings, the service orientation can be
achieved.
• Architecture Building: This second phase is the key to deliver
service orientation into all enterprise operations and systems, in
order to be more agile and efficient, through its strategic partner-
ship of implementation platform. This involves a dedicated team
designed for SOA that can collaborate between all stakehold-
ers to design, develop and govern the entire service orientation
initiative implementation that includes service components and
infrastructure.
• Services Operation: This third phase is the core of the ser-
vice orientation initiative: the operating of loosely coupling new,
reusable and existing software systems and services in a single
integration system of the legacy architecture. By referring to the
operational guideline, enterprise can ensure the service orienta-
tion works accordingly and quickly reacts to necessary business
change inline with the change management plan.
• Continuous Improvement: This fourth phase is at the point
where service orientation continuously improves the full range
of services in order to satisfy the requirements of core business
processes. It also includes in which the performance of service
orientation is measured and updated across the enterprise for
effectiveness.
From this research, we conclude that the general four-phase
service orientation initiative process derived from the literature
and has been affirmed by the case study is a suitable process to
implement the service orientation initiative in leading to becom-
ing a full services environmental enterprise.
Acknowledgments: This research is funded by the Uni-
versiti Teknikal Malaysia Melaka (UTeM) and we gratefully
acknowledge its generous financial support.
References and Notes1. M. P. MacDonald and D. Aron, Leading in Times of Transition: The 2010 CIO
Agenda, Gartner Inc., Report (2010).2. Y. Gong and M. Janssen, Government Information Quarterly (2011).3. J. Choi, D. L. Nazareth, and H. K. Jain, Journal of Management Information
Systems 26, 253 (2010).4. S. Bistarelli and P. Campli, Fairness as a QoS measure for web ser-
vices, Young Researchers Workshop on Service-Oriented Computing (2009),pp. 115–127.
5. P. Bianco, R. Kotermanski, and P. Merson, Evaluating a serviceoriented archi-tecture, Software Engineering Institute, Technical Report (2007).
6. H. M. Chen, R. Kazman, and O. Perry, IEEE Transactions on ServicesComputing 3, 145 (2010).
7. T. Kokko, J. Antikainen, and T. Systä, Adopting SOA—Experiences from nineFinnish organizations, European Conference on Software Maintenance andReengineering (2009), pp. 129–138.
8. G. Feuerlicht and S. Govardhan, SOA: Trends and directions, Proceedingsof 17th International Conference on Systems Integration, Czech Rep. (2009),pp. 149–155.
9. M. S. Sulong, A. Koronios, J. Gao, and A. Abdul-Aziz, Implementing qual-ity service-oriented architecture initiative in organisations, Portland Interna-tional Conference for Management of Engineering and Technology (2012),pp. 3666–3673.
10. B. Hensle, Oracle’s approach to SOA, Oracle Corporation, IT Strategies fromOracle: Data Sheet (2010).
11. A. Arsanjani and K. Holley, Increase flexibility with the service integra-tion maturity model (SIMM): Maturity, adoption, and transformation to SOA.IBM.com (2005).
12. R. Altman, SOA overview and guide to SOA research. Gartner Inc., Report(2010).
13. T. C. Shan and W. W. Hua, Service-oriented architecture roadmapping,Congress on Services-I, CA, USA (2009), pp. 475–476.
14. A. Moeini, N. Modiri, and T. Azadi, Service oriented architecture adoptionmanagement roadmap, 7th International Conference on Digital Content, Mul-timedia Technology and its Applications, Iran (2011), pp. 119–124.
15. K. M. Eisenhardt, Academy of Management Review 14, 532 (1989).16. K. M. Eisenhardt and M. E. Graebner, Academy of Management Journal
50, 25 (2007).17. A. Fontana and J. H. Frey, The interview: From structured questions to negoti-
ated text, Handbook of Qualitative Research, 2nd edn., edited by N. K. Denzinand Y. S. Lincoln, Sage Publications, CA (2000), p. 645.
18. A. Lewins and C. Silver, Using Software in Qualitative Research: A Step-by-Step Guide, Sage Publication, USA (2007).
Received: 25 November 2014. Accepted: 17 January 2015.
Nurbaya Mohd Rosli1, Nurazean Maarop2�∗, and Ganthan Narayana Samy2
1Kolej Vokasional ERT Setapak, Kuala Lumpur, Malaysia2Advanced Informatics School, Universiti Teknologi Malaysia, Jalan Yahya Petra, 54100 Kuala Lumpur, Malaysia
E-learning has been implemented by education institution worldwide for decades, including Kolej VokasionalERT Setapak, under the management of Ministry of Education, Malaysia. Implementing a successful e-learningnamely FROG VLE, is another step of collaborating education with technology, thus, it is important for collegeto recognize the acceptance factor of FROG VLE among teachers. This study aims to explore the acceptancefactors namely, perceived usefulness, perceived ease of use, computer self-efficacy, convenience, champion’scharacteristics, instructional design, and technological factor towards behavioral intention as technology accep-tance indicator. The research is based on mix-method study involving semi-structured interviews and question-naires survey. All these factors were explored to examine factors that influence the acceptance. The resultsindicated that perceived usefulness, perceived ease of use, instructional design, convenience, technologicalfactor and computer self-efficacy have significant effect to the acceptance of FROG VLE. However, champion’scharacteristic was found quantitatively insignificant towards the acceptance of FROG VLE but bear relevantto be considered as important from qualitative standpoint. Finally, providing e-learning technology acceptancefactors may help education institution to implement and enhance the technology for efficient use of e-learning.
Keywords: Technology Acceptance, E-Learning, School Online Learning.
1. INTRODUCTIONDevelopment in Information and Communication Technologies
(ICT) has its impact on education sector in Malaysia. E-learning
is an innovative approach for delivering electronically mediated,
well-designed, student-directed and interactive learning environ-
ment for everyone, regardless of time and place, using either the
Internet or digital technologies in collaboration by the principles
of instructional design.1 In Malaysia, Learning Management Sys-
tem (LMS), an e-learning, has been implemented by pilot project
since 2009 using the open source technology, LMS Moodle.2
E-learning is defined as learning facilitated and supported
through the utilization of information and communication tech-
nologies (ICTs).3 Effective implementation of an e-learning
initiative requires attention to a number of issues including tech-
nological, pedagogical and individual factors.4 Besides, good
connectivity is very important in order to realize the functionality
of e-learning used by the community.
In 2010, through Bahagian Teknologi Pendidikan (BTP),
Ministry of Education (MOE) has implemented an e-learning;
“Learning Management System” (LMS) to 50 pilot schools
throughout Malaysia. According to Assistant Director, Learning
∗Author to whom correspondence should be addressed.
Resources Sector of BTP, the constraints when implementing
LMS are the internet connection problem and users were com-
plaining on the interface of the e-learning. There are eight out
of 50 pilot schools do not use LMS in implementation within
10 months, likely the LMS system in the school is not con-
nected with BTP server.2 In 2012, a new e-learning system
known as FROG VLE was introduced. The system is a cloud-
based learning platform that can be accessed by teachers, stu-
dents and parents. This project is currently been implement in all
school in the country by using the 4G internet technology under
1BestariNet program.
In developing countries, technological factors still remain as
an obstacle in implementing online learning system, where the
advancement of IT infrastructures development countries is far
behind from their developed counterparts.5�6 Therefore, this study
explores the acceptance using appropriate acceptance model.
In regard to the context of this study, 1BestariNet is an “End
To End” (E2E) network service for the purpose of teaching
and learning process and for management and administration
of all 10,000 schools under Ministry of Education, Malaysia.
This project is the main platform for the proposed e-learning
concept (FROG VLE). The implementation of this project is to
enhance the internet access bandwidth for all school and the
ment sharing systems and learning resources.7�8 VLEs are rapidly
becoming an integral part of the teaching and learning process.8�9
Thus, it enables improvements in communication efficiency, both
between student and teacher, as well as among students.7�8
User acceptance has been viewed as the pivotal factor in deter-
mining the success or failure of any information system project.10
Most of this work been focused on productivity-oriented or util-
itarian systems.11 Therefore, user acceptance is defined as will-
ingness to employ IT for the tasks it is designed to support.12
Throughout the study, there are several models have been devel-
oped to investigate and understand the factors involving the
acceptance of computer technology in organizations. The models
are about to study the user acceptance, adoption and IT usage
behaviour.
Virtual Learning Environment (VLE) is known as a good plat-
form of e-learning, specifically in this study for FROG. The study
from various journals show many researchers have been review-
ing on the acceptance and adoption of e-learning using different
types of technology acceptance model. The studies used the com-
bination of keywords in searching the acceptance theories and
related e-learning acceptance. Most of search is from Science-
Direct and IEEE databases.
From the literature review of e-learning acceptance, it
shows that most researchers are using Technology Acceptance
Model (TAM)13 as their research model. According to William
and Jun,14 TAM is a powerful and robust predictive model which
results from the meta-analysis of technology acceptance model
study.
One of the researches in the e-learning acceptance area was
done by Hussein et al.5 as shown in Figure 1 and they used TAM
as their base theoretical model. The objective of their study was
to investigate the factors that affect acceptance of e-learning. This
research was conducted using quantitative methodology where
the respondents were online students of the Indonesian Open
University (UT) from various major of study. As a result, this
research had support instructional design, computer self-efficacy
and technological factor as predictors of e-learning acceptance in
a developing country. According to Šumak,15 TAM has been used
significantly to explain technology acceptance in schools either
among students or teachers. Hence, this research has considered
appropriate model to help explain the acceptance of FROG VLE
in the context of the study.
InstructionalDesign
ComputerSelf Efficacy
TechnologicalFactor
Convenience
Instructor’sCharacteristic
PerceivedUsefulness
PerceivedEase of USe
Intention toUse
Fig. 1. Original model by Hussein et al.5
In regard to the current environment of FROG VLE implemen-
tation, there was no presence of instructor in the environment.
However, e-learning champion had been selected to assist and
share their skill and knowledge. Hence, instructor characteristic
from the original model as shown in Figure 1 has been replaced
with champion’s characteristics as an external factor in this study
context.
Self-efficacy is defined as the judgment of one’s capability to
use an IT or computer.17�18 Previous studies have shown that
computer self-efficacy is related to technology acceptance.5�18
The design of online learning is similar type of classroom
format where there are course description, objectives, content,
purpose, scope and evaluation.19 According to Hussein et al.5
a well-designed application is believed to have an effect on online
learning adoption. Therefore, teachers must be able to easily find,
read, download and save the materials that been shared online.
Technological factors still remain as an obstacle in imple-
menting online learning system in developing countries, where
the advancement of IT infrastructures is far behind from their
developed counterparts.20 According to Peters (2002),21 problems
with connection, low modem speed, missing links, loading page
and availability of memory are some of the obstacles in online
learning.
Convenience is one of the enabling factors identified in the
online learning literature [5]. According to Tobin (1998),22 con-
venience is achieved when students can access the learning at
convenience time. For this study, teachers can implement a class
whenever and wherever they want, as the students will access
accordingly.
Champion’s characteristic is another factor believed to influ-
ence teacher’s acceptance of FROG VLE. A successful imple-
mentation of e-learning does not only rely on advanced
technology, but also champions as a reference key people, who
is responsible to share their knowledge and encourage teachers
to use FROG VLE. According to a study in Australia in 2009,22
champion can be regarded as an agent that can set the learning
of an e-learning where champions is skilled in e-learning, willing
to share their expertise with passion and enthusiasm and willing
to solve problems, either technical or non-technical problems.
Their study discovered that e-learning champions have the abil-
ity to empower, motivate and mentor teachers in e-learning and
establish effective networking with other teachers to encourage
knowledge transfer and technology exploration.22 Figure 1 illus-
trates the altered research model by Hussein et al.5 considering
3374
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3373–3377, 2015
champion characteristics whereas the overall proposed model to
be used in this study was shown in Figure 2.
In summary, the theoretical background of this study is
based on Technology Acceptance Model with external factors by
Hussein et al.,5 where the research evidently support the instruc-
tional design, technological factors and computer self-efficacy
factors as a very important role in acceptance towards e-learning.
However, convenience and instructor’s characteristics were not
dominant factors in the research, therefore, for this study, cham-
pion’s characteristic was replacing instructor’s characteristics.
Thus this is in line with the recommendation from the previous
study to look into other factors which were not addressed by
their.
3. METHODOLOGYThe study used mixed-method research design.23 In particular,
in gathering the data, a mixed-method convergent parallel design
introduced by Creswell and Clerk (2011)23 was implemented to
compare or relate the data and followed by interpretation from
the result gained. Questionnaires and interview questions are pre-
pared to comply with the studies. The purpose of this approach
was that quantitative data and subsequent appropriate analyses
would provide a strong understanding to the whole results.23�24
Five teachers participated in interview whereby thirty teachers
representing the 83% population of FROG VLE participated
in the questionnaire survey. In designing the questionnaire, the
study used the five likert scale questions to obtain the level of
agreement or disagreement. Seven out of eight factors used in
the survey for this study were adapted from the previous study
and one factor is included for the context of this study.
4. RESULTIn regard to quantitative analysis, descriptive analysis and the
Spearman correlation were applied for quantitative analysis.
On the other hand, the qualitative findings of this study were
based on thematic analysis and the presentation of findings
were based on weighting-the-evidence approach by Miles and
Huberman.25 The results of the study are as follows.
4.1. Quantitative Result
Table I shows the correlation coefficient analysis result for the
study. From the table, it illustrates the association between BI
and all variables except CC were considerably strong (rs > 0�50),
while CC was not significant as the p-value is more than 0.05.
PerceivedUsefulness
ComputerSelf Efficacy
PerceivedEase of Use
Intentionto Use
InstructionalDesign
TechnologicalFactor
Convenience
Instructors’Characteristic
Fig. 2. Proposed model.
Table I. Correlations of factors.
Correlation strengthBI and significance
BICorrelation 1.000 –Sig. (2 tailed) 0.000
PUCorrelation 0.794 Strong and significantSig. (2 tailed) 0.000
PEOUCorrelation 0.751 Strong and significantSig. (2 tailed) 0.000
CSECorrelation 0.507 Strong and significantSig. (2 tailed) 0.004
CCCorrelation 0.417 Not significantSig. (2 tailed) 0.022
CONVCorrelation 0.801 Strong and significantSig. (2 tailed) 0.000
TFCorrelation 0.674 Strong and significantSig. (2 tailed) 0.000
IDCorrelation 0.784 Strong and significantSig. (2 tailed) 0.000
4.2. Qualitative Result
The findings of the study of perceived usefulness indicated that
this factor shows that FROG VLE can improves and enhance
teaching and learning activities. Majority of the respondent stated
about the positive impact when they used FROG VLE during
their teaching process. Example of excerpts:
“� � �FROG VLE helps in my teaching performance,where I can use notes that been share with otherteachers and I can surf any new materials throughfrog sites and frog store � � �” �Participant 1�
“� � � it improves my quality of teaching, from tra-ditional style to the use of ICT especially for myteaching materials � � �” �Participant 5�
In this study for variable perceived ease of use, all respon-
dents represent that perceive ease of use supports the use of
FROG VLE. Respondent 5 finds FROG VLE is quite easy, but
need more time to understand the e-learning functions, especially
the widgets and emphasis that the lack of exposure and using
online applications affects the understanding of using FROG
VLE. Example of excerpts:
“� � � It is quite easy for me, but I need sometimesto understand the functions especially the use ofwidget � � �” �Participant 3�
“� � �The widget ease me, just drag from the menu.Plus, I use Google drive to import material intoFROG � � �” �Participant 1�
“� � � It is easy. More or less is similar to FBfunctions � � �” �Participant 2�
The majority of the respondents show that they are IT savvy
user. Most of them are exposed to use computer in their work,
specifically in teaching. However, based on counting measure-
ment, From the descriptive analysis, the majority of the users
3375
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3373–3377, 2015
have ticked for “Neutral” in computer self-efficacy, this conclude
that, teachers in KVertS need guidance to use FROG VLE. Exam-
ple of excerpts:
“� � �No. At the first time, I need someone who skill-ful to use it to assist me � � �” �Participant 3�
“� � �Generally, I used to download and save filefrom online site, so I don’t have problem on it.I’m confident � � �” �Participant 4�
“� � �No. I seldom use application online, soI need someone to assist me in using FROG � � �”�Participant 5�
During the interview, interviewee were asked about whether
FROG champion’s in the college really help in exploring and
use the e-learning, the results were exactly the same. The FROG
champions are skillful and willing to share their expertise,
besides they are friendly and approachable. Example of excerpts:
“� � � the champion is very good. He is very help-ful and eager to guide me to use FROG � � �”
�Participant 5�
“� � �They are eager to share whatever they know.If I asked them via phone, they still giveguidance � � �”
In regard to technological factor, all respondents emphasized
that, poor internet connections demotivate them to use FROG
VLE in their teaching activity.
“� � �The interface is good and user friendly. If theinternet connection is poor, it gives problem � � �”�Participant 1�
“� � � Interface is fantastic and user friendly. Withoutgood internet connection, it takes so long for thepage to load. This is frustrating � � �” (Participant 2)
4.3. Overall Mixed-Method Finding
The results of both interviews and questionnaires analysis are
shown in Table II.
In regard to Perceived Usefulness, the results of both data are
consistent. Teachers agreed that (PU) affects their acceptance of
FROG VLE.
In regard to Perceived Ease of Use, The results of both data are
consistent. Teachers agreed that (PEOU) affects their acceptance
of FROG VLE.
In regard to Computer Self-Efficacy, the results of both data are
consistent. Teachers agreed that (CSE) affects their acceptance
of FROG VLE.
In regard to Champion’s Characteristics, the results of both
data are not consistent. Spearman’s correlation result is not sig-
nificant for (CC).
In regard to Convenience, the results of both data are consis-
tent. Teachers agreed that (CONV) affects teachers’ acceptance
of FROG VLE.
In regard to Technological Factors, the results of both data are
consistent. Teachers agree that (TF) affects teachers’ acceptance
of FROG VLE. Even though the current state of technology provi-
sion is not favorable but the respondents agreed TF is very impor-
tant for usage continuity. Further, the correlation indicates strong
relationship between TF and the acceptance of FROG VLE.
Table II. Concluding result based on mixed-method.
Interview Questionnaire ConcludingFactors responses results result
Perceivedusefulness
StronglyRelevant
Significant correlationwith 0.794
Significant.
Perceived easeof use
Relevant Significant correlationwith 0.751
Significant.
Computerself-efficacy
Relevant Significant correlationwith 0.507
Significant.
Champions’characteris-tics
StronglyRelevant
Not significantcorrelation with0.417; sig. (2-tailed)at 0.022
Less significant.
Convenience StronglyRelevant
Significant correlationwith 0.801
Significant.
Technologicalfactors
StronglyRelevant
Significant correlationwith 0.674
Significant.
Instructionaldesign
StronglyRelevant
Significant correlationwith 0.784
Significant.
Behavioralintention
StronglyRelevant
Significant value forall factors thatassociate with (BI)except (CC)
Significant.
In regard to Instructional Design, the results of both data are
consistent. Teachers agree that (ID) affects teachers’ acceptance
of FROG VLE.
Overall, the FROG VLE is accepted by teachers in KVertS as
indicated by Behavioral Intention.
5. CONCLUSION AND DISCUSSIONThis study has identified factors influencing the acceptance of
education technology in the vocational school environment. The
study has further explored the acceptance of FROG VLE using
the proposed model. In order to obtain more beneficial results in
future research, this study suggests the following aspects.
• Further researches may extend the scope of study by including
samples from different educational institution such as primary
school or secondary school.
• Include more elements and factors that affect the acceptance
of FROG VLE among teachers in the research to achieve more
understanding regarding FROG VLE acceptance.
• Use other acceptance models to explore the acceptance of
FROG VLE among teachers.
• Apply observation method for data collection to gain more
results for the research.
Findings obtained from this research shows that perceive ease
of use, computer self-efficacy, convenience, instructional design,
champion’s characteristics and technological factors have sig-
nificant effect towards the acceptance of FROG VLE among
teachers. From this result, educational institution may have more
insight in accepting and implementing e-learning in college or
school for more effective teaching and learning activity.
References and Notes1. J. L. Morrison and B. H. Khan, The global e-learning framework: An inter-
view with Badrul Khan, The Technology Source, A Publication of the MichiganVirtual University (2003).
2. T. S. Wai, Jurnal Pembestarian Sekolah 2010 BTP KPM (2010).3. M. Jenkins and J. Hanson, E-Learning Series: A guide for Senior Managers,
Learning and Teaching Support Network (LTSN) Generic Centre, UnitedKingdom (2003).
3376
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3373–3377, 2015
4. M. Masrom and R. Hussein, User Acceptance of Information Technol-ogy: Understanding Theories and Models, Venton Publishing, Kuala Lumpur(2008).
5. R. Hussein, U. Aditiawarman, and N. Mohamed, E-learning acceptance ina developing country: A case of the Indonesian Open University, GermanE-Science Conference (2007).
6. W. C. Poon, L. T. Low, and G. F. Yong, The International Journal of EducationalManagement 18, 374 (2004).
7. L. L. Martins and F. W. Kellermanns, Academy of Management Learning andEducation 3, 7 (2004).
8. J. J. L. Schepers and M. G. M. Wetzels, Technology acceptance: A meta-analytical view on subjective norm, Proceedings of the 35th European Mar-keting Academy Conference, Athens, Greece (2006).
9. K. A. Pituch and Y. K. Lee, Computers and Education 47, 222 (2006).10. F. D. Davis, A Technology Acceptance Model for Empirically Testing New End-
User Information Systems: Theory and Results, Sloan School of Manage-ment, Massachusetts Institute of Technology: Doctoral Dissertation (1986).
11. V. Venkatesh and F. D. Davis, Management Science 46, 186 (2000); F. D.Davis, MIS Quarterly 13, 319 (1989).
12. A. Dillon and M. Morris, User acceptance of new information technology:Theories and models, Annual Review of Information Science and Technol-ogy, edited by M. Williams, Information Today, Medford, NJ (1996), Vol. 31,pp. 3–32.
13. F. D. Davis, MIS Quarterly 13, 319 (1989).14. R. William and H. Jun, Information and Management 43, 740 (2006).
15. B. Šumak, M. Hericko, M. Pušnik, and G. Polancic, Informatica 35, 91(2011).
16. R. Agarwal and E. Karahanna, MIS Quarterly 24, 665 (2000).17. M. E. Gist, Personnel Psychology 42, 787 (1989).18. E. E. Grandon, K. Alshare, and O. Kwun, Journal of Computing Science in
College 20, 46 (2006).19. W. C. Poon, L. T. Low, and G. F. Yong, The International Journal of Educational
Management 18, 374 (2004).20. O. Peters, Distance Education in Transition: New Trends and Challenges,
Bibliotheks-und Informationssystem der Universität Oldenburg, Oldenburg(2002), pp. 37–45.
21. K. Tobin, Qualitative perceptions of learning environment on the world wideweb, International Handbook of Science Education, edited by B. J. Fraserand K. G. Tobins, Kluwer Academic Publishers, United Kingdon (1998),pp. 139–162.
22. M. Jolly, B. Shaw, K. Bowman, and C. McCulloch, Final report—The impactof e-learning champions on embedding e-learning—In organisations, indus-try or communities, Department of Education, Employment and WorkplaceRelations, Australian Government (2009).
23. J. W. Creswell and V. L. P. Clark, Designing and Conducting Mixed MethodsResearch, Sage, Thousand Oaks, CA (2011).
24. N. Maarop and K. T. Win, Journal of Medical Systems (J. Med. Syst.) 36, 2881(2012).
25. M. B. Miles, A. M. Huberman, and J. Saldana, Qualitative Data Analysis:A Methods Source Book, Sage Publication (2013).
Received: 27 November 2014. Accepted: 23 January 2015.
Najihahbinti Mustaffa∗, Zalehabinti Ismail, Zaidatunbinti Tasir, and Mohd Nihra Haruzuan Bin Mohamad Said
Faculty of Education, University Technology Malaysia, 81310, Skudai, Johor, Malaysia
Algebra is interrelated with other mathematical topics such as statistics and geometry. Learning algebra is aboutsolving an unknown and is the way of thinking. Algebraic thinking is a process of thinking in specific domain inMathematics. Teaching and learning process of algebra should be reform nowadays as to beyond computationaland assisted with technology. This paper is presented that algebraic thinking should be developed throughtechnology concurrently with the role of teacher. The study of developing algebraic thinking through technologyintegrated with the role of teacher should be conducted.
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3382–3384, 2015
Table I. List of published researches on technology.
Author Software Result Level
(Baltaci and Yildiz, 2015) Geogebra 5.0 version Program is usable. Facilitate learning aboutthree-dimensional objects. Can help the acquisitionof age-dependant cognitive level of learners tounderstand three-dimensional objects at an earlierage. Use their time effectively using Geogebra
Pre-service elementaryteacher
(Tabach et al., 2013) Spreadsheet Understanding of the situation and ease of access tothis generational activity. Shows flexible use ofexpressions in a spreadsheet environment from thevery beginning of the school year. Able to generateonly multi-variable and recursive expressions. CIEapproach to beginning algebra can facilitate thetransition from arithmetic to algebra
Primary school students
(Hall and Chamblee, 2013) Geogebra User friendly. Understand the concept of geometricconcept, transformation, and rotation. Enable tosee the changes of equation when the parabolamoves around. Show the changes result onequations
Pre-service teacher
(Bokhove and Drijvers, 2012) MLwiN 2.22 with estimationprocedure RIGLS
Improvement in score Improvement to recognizepatterns Have a sense for symbols
Middle school students
(Bills, Wilson, and Ainley, 2010) Spreadsheet Able to construct meaning for variable. Able toformalize the expression of functional relationships.No kind of thinking that pupils engage during ateaching programme
Primary school students
(Zeller and Barzel, 2010) CAS Influence the student that algebra related toarithmetic Influences the student to evaluatealgebraic representation
Middle school students
(Pierce et al., 2009) CAS CAS worth to be used by most of the students Highperformance students able to manipulate algebraicexpressions Low performance students faceddifficulties to understand the algebraic symbolsand structure of CAS syntax
Middle secondary schoolstudents
the role of a teacher. The significance of the role of a teacher
that is able to ask questions in enhancing students’ thinking is
important. Reversibility, flexibility and generalization of ques-
tions able to develop the generalization as well as deepen stu-
dents’ thinking.6 These types of questions are important as the
questions will lead students to answer, create problems, develop
multiple ways of solution methods, and able to predict the
answers or check the responses from the student. In the learn-
ing environment, students will learn by thinking and not just
memorizing. Students developed their algebraic thinking by gen-
eralizing and relating the problems through sophisticated ways
of thinking such as organizing and manipulation.17 Teachers not
only play a significant role in enhancing students’ thinking, but
technology is also pivotal in this focus.
Nowadays, technology has been used extensively and
widely in learning environment. Algebraic thinking may be
developed using computer environments. Based on previous
studies,3�4�6�8�10�14�15�19 various software have been used to
develop algebraic thinking. Students should be able to understand
the symbols and operations of technology, algebra symbols,4 and
most importantly, the problem of transition between arithmetic
and algebra. CAS can solve the problem of transition between
arithmetic and algebra,19 although it is costly and time consum-
ing. Spreadsheet is applicable only for the basic of generalization.
Furthermore, a previous study5 mentions that for the primary
school students, the thinking engaged in the process of learning
is non-existent. Technology serves as a scaffold in a learning pro-
cess thusit should be efficient, effective, and able to save time in
a learning process such as in learning Geogebra. It must also be
able to improve the view of three-dimensional objects in the early
ages, although this should be tested on specific characteristic of
algebraic thinking.
Computers will help students to connect formal representa-
tions with dynamic visual representations, and allow students
to explore, express and formalize informal ideas. However, the
importance of computers in the learning environment is similar
to the teachers’ role in the classroom. Algebraic thinking requires
problem-solving supported with types of questioning and tech-
nology concurrently.
5. DISCUSSIONThe authors expect that in learning activities of technology, alge-
braic thinking can affect student engagement. In any event, its
intended effect may cause positive or negative impacts. The use
of technology has the potential to be implemented in developing
algebraic thinking, but it will invite challenges and constraints
as mentioned. Viewed from several angles, the authors propose
some alternatives to overcome the constraints and challenges.
Nevertheless, the study on the effectiveness of algebraic think-
ing through technology is still ongoing, and that it may yield dif-
ferent results from those concluded in previous studies. Through
this research, the authors will unlock any issues and constraints
that arise.
6. FUTURE STUDIESThe potential study of algebraic thinking through computers
should be explored. Algebraic thinking software should be
3383
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3382–3384, 2015
extended to explore the relationships, manipulation of symbols
and procedures, reasoning with representations, using algebra as
a tool and connecting representation. Most of the findings empha-
sized on the generalization with algebraic symbols which are
in generational and meta-level activities, although the study of
transformational activities has yet to be explored.15
The authors would like to know the implication of developing
algebraic thinking through technology concurrently the role of a
teacher. Although there are many benefits of using computers,
teachers also offer many advantages as well.
7. CONCLUSIONNowadays, computers are available to be used in the learning
environment.15 ICT tools allow students to use different strate-
gies through a stepwise approach with procedural skills as well
to enhance their conceptual understanding.6 Therefore, we can
conclude that the necessity of ICT tools in learning mathemat-
ics, particularly in developing algebraic thinking. However, the
role of teacher is still needed even though the ICT tools are
widely used. The teacher also should complete themselves with
the knowledge of content, ICT skills, and skills of questioning.
References and Notes1. O. S. Abonyi and I. Nweke, Journal of Education and Practice 5, 31 (2014).2. O. A. Alghtani and N. A. Abdulhamied, Procedia—Social and Behavioral
Sciences 8, 256 (2010).
3. S. Baltaci and A. Yildiz, Cypriot Journal of Educational Sciences 10, 12(2015).
4. R. Banerjee, Contemporary Education Dialogue 8, 137 (2011).5. L. Bills, K. Wilson, and J. Ainley, Research in Mathematics Education 7, 67
(2010).6. C. Bokhove and P. Drijvers, Computers and Education 58, 197 (2012).7. G. Booker and W. Windsor, Procedia—Social and Behavioral Sciences 8, 411
(2010).8. B. Dougherty, D. P. Bryant, B. R. Bryant, R. L. Darrough, and K. H.
Pfannenstiel, Intervention in School and Clinic 1 (2014), Doi: 10.1177/1053451214560892.
9. S. Gutiérrez, M. Mavrikis, and D. Pearce, A learning environment for promot-ing structured algebraic thinking in children, Eighth IEEE International Con-ference on Advanced Learning Technologies (2008), pp. 74–76.
10. J. Hall and G. Chamblee, Computers in the Schools: Interdisciplinary Journalof Practice, Theory, and Applied Reseacrh 30, 12 (2013).
11. C. Kieran, The Mathematics Educator 8, 139 (2004).12. J. Lee and J. Pang, Journal of the Korea Society of Mathematical Education
16, 2070 (2012).13. B. Patton and E. D. L. Santos, International Journal of Instruction 5, 5
(2012).14. R. Pierce, L. Ball, and K. Stacey, International Journal of Science and
Mathematics Education 7, 1149 (2009).15. M. Tabach, R. Hershkowitz, and T. Dreyfus, ZDM-International Journal on
Mathematics Education 45, 377 (2013).16. W. Windsor, Algebraic ThinkingD: A problem solving approach, Proceedings
of the 33rd Annual Conference of the Mathematics Education Research Groupof Australasia (2008), pp. 665–672.
17. W. Windsor, APMC 16, 8 (2011).18. W. Windsor and S. Norton, Developing algebraic ThinkingD: Using a problem
solving approach in a primary school context, MathematicsD: Traditions and[New]Practices (2011), pp. 813–820.
19. M. Zeller and B. Barzel, ZDM—International Journal on MathematicsEducation 42, 775 (2010).
Received: 29 November 2014. Accepted: 25 January 2015.
Azlina A. Rahman∗, Baharuddin Aris, Mohd Shafie Rosli, Hasnah Mohamed,Zaleha Abdullah, and Norasykin Mohd Zaid
Department of Educational Science, Mathematics and Creative Multimedia, Faculty of Education,Universiti Teknologi Malaysia, Johor Bahru 81310, Malaysia
Flipped classroom is the latest pedagogy that has grown across multi-discipline and age levels which effec-tiveness has been proven empirically. The flipped classroom approach allows the students to review the topicsgiven prior to learning them in the classroom and apply the knowledge gained practically via in-class activities.Therefore, students are given more opportunities to apply the knowledge they have learned into the real lifesituation in collaborative learning environment. Furthermore, flipped classroom shows most impactful effects onincreasing learning interactions, improving students’ achievement and boosting critical thinking. Studies haveindicated that flipped classroom could also stimulate students’ interest and could even improve their attitudestowards school. Students are able to receive a personalized education to suit their learning style while syllabuscould be covered before time. However, the key feature of successful criteria in flipped classroom is students’preparedness. Very few of the reviewed articles emphasizes on this critical aspect. The need for the students tobe prepared prior to the teaching and learning process plays an important role to make this approach successfuland meaningful. This is because if the students come to the class unprepared, they will give a blank look andwill not get involved in the classroom. The school is a place to improve working improvement and to producestudents with maximum academic growth. While the flipped classroom approach has been seen successfulfrom the perspectives of both students and teachers, the authors notice that, there is still room for improvementin some areas. The authors choose to redesign the approach by factoring in the preparedness aspect. Thispaper summarizes the importance of preparedness based on limited past researches and also presents somepossible ways of redesigning the prior learning process in flipped classroom for secondary education.
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3389–3391, 2015
impact on students’ performance, both at the tertiary9�16–23 and
the school levels.12�13�15�25�26
2. THE KEY FEATURE IN FLIPPED
CLASSROOMAlthough there have been many studies regarding the flipped
classroom in various dimensions and disciplines, those which
emphasize on students’ preparedness for the flipped classroom are
still lacking. The authors see that the key feature in the flipped
classroom is preparedness because if students are not prepared in
the first phase of the flipped classroom learning process, flipped
classroom learning objectives will not be achieved. Taking into
account some aspects to fulfill learning requirements of the stu-
dent, especially those in the secondary schools, the authors pro-
pose preparedness as one of the initiatives to be made available
in the flipped classroom learning process. Preparedness refers to
the stage of readiness of the students in the flipped classroom
approach.
The original flipped classroom learning process comprises two
learning phases.1 The first learning phase is a self-paced learning,
where teachers provide assignments for students to read about
what will be studied in the next lesson. Hence, during the flipped
classroom second phase session, the teachers could allocate more
time and provide more opportunities for their students. Interac-
tions between students and teachers could be improved. There-
fore, the teachers would find it easier to identify any students’
misunderstanding early. The students also learn from the teachers
and their peers who are more competent to promote the devel-
opment of knowledge and soft skills of the students. There are
previous studies26 which have made modifications on the flipped
classroom as shown in Figure 1.
Summarize the harvestpropose confusion
Learning by one selfSelf paced
Exhibit communicationResearching cooperatively
Scientific experimentAccomplish Homework
Layout preview
Out
of c
lass
In th
e cl
assr
oom
Fig. 1. Teaching structure.26
Redesign Phase: Prior to teaching andlearning process
Option 1:Students are given questions about what
is read in Phase 1. This may be done either online or offlineor both.
Option 2:Students perform instantaneous activitiesrelated to what has been learnt. It couldbe as in the written form such as quizzes
and education game.
Option 3:Each student is to ask questions to the
teachers or peers.
Phase 2
Activities in the classroom such as groupdiscussion, demonstrations, lab activities
and debate.
Phase 1
Students are exposed to read or referencematerials through the use of digital
technologies such as video and the InternetO
ut o
f cla
ssIn
the
clas
sroo
m
Fig. 2. Redesign stage of flipped classroom preparedness.
According to Figure 1, the flipped classroom learning process
is divided into two phases which are learning in the class-
room and learning outside the classroom. Learning in the class-
room means self paced learning while learning outside the
classroom consists of hands-on learning. However, the teaching
structure shown in Figure 1 is feasible at the tertiary level26 has
developed the activities as portrayed in the dashed box. In the
context of this study, the authors see the potentials if the steps
in the dashed box could be applied in various levels, particularly
at the school level. The authors feel that there is some looseness
in Figure 1, especially in the dashed box, namely, what if the
students do not make the task as directed by the teachers. If the
students are not ready, they will come to class unprepared with
a blank look. This situation will potentially become worse as the
students are not able to fully absorb the lesson which will take
place. It is to cater for this need that, the author has drawn up a
guideline in the flipped classroom learning process, as depicted in
Figure 2. These guideline is strongly suggested by Refs. [27, 28]
that the teacher or instructor must also develop activities and/or
pretest to ensure that students are prepared for class.
In reference to Figure 2, the original version of flipped class-
room consists of two phases, namely Phase 1 (out of class) and
Phase 2 (in the classroom) as shown in Figure 1. As the author
3390
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3389–3391, 2015
sees the importance for preparedness in the flipped classroom
to be emphasized, the author has redesigned a phase marked in
the dashed box in between Phase 1 and Phase 2 in Figure 2.
Although there are researchers21�29 who are concerned about the
preparedness such with emphasis on online learning using Intelli-
gent Tutoring System as one of the initiatives to ensure that each
student is interactive with what has been learnt, the implementa-
tion of such an approach is limited only to the university students.
At the school level, only researchers12 have made it compulsory
for each student to submit questions before the class starts to
ensure that they are ready for the assignment to be given. It is
obvious that, studies which emphasize on the preparedness aspect
of the flipped classroom research are very limited. Findings in a
research by Refs. [27–29] prove that students’ preparedness with
regard to digital materials or assignment given prior to teaching
and learning results in active learning, compared to students who
are not prepared with the provided materials.
The authors see redesigning the flipped classroom approach
as a measure that needs to be done so that students are more
prepared for the assigned learning materials. Using the mea-
sures provided, it will assist teachers to ensure the success of
the flipped classroom approach. The redesigned method provides
teachers with the guidelines to implement the steps shown before
going to Phase 2 in Figure 2. Based on Figure 2, for schools
with adequate infrastructure facilities such as Internet and com-
puter hardware, Option 1 in the redesigned phase does not pose a
problem. However, for schools with problems in terms of infras-
tructure, they may opt for off line learning. For teachers who have
the time and creativity, they may capitalize on this opportunity
by giving written exercises such as quizzes and education games
as proposed in Option 2. The simplest approach is Option 3,
in which, the teachers require each student to ask questions and
the discussion sessions are continued in the following phase as
made by researchers.12
Comparing Figure 1 with Figure 2, especially in the dashed
boxes, the authors emphasize the preparedness aspect, without
which, the objectives of the flipped classroom approach are hard
to be achieved. This situation could occur not only at the school
level but also at the tertiary level. In the context of this study, the
authors chose an Option 1 to be applied outside the classroom
as a self-pace learning.
3. CONCLUSIONOverall, flipped classroom has a positive impact on students’
achievement. However, when viewed in terms of infrastructure,
it should be given a little touch to make it more feasible.
The authors see the need for this issue to be highlighted with
the redesigned approach with the introduction of preparedness
aspect, especially in the secondary education. Secondary school
students need to be kept motivated to make learning more inter-
esting and meaningful. The proposal by the authors are also
strongly suggested by Refs. [27–29] who also emphasized on the
importance of students’ preparedness before the flipped class-
room teaching and learning begins. Although the researcher sug-
gested a preparedness phase in applying the flipped classroom,
the use of online and offline technology also helps in improving
students’ achievement.12�14�20–26�30 However, the survey data can-
not be reported since the data collection process is still ongoing.
Acknowledgments: The authors would like to honour
the continuous support and encouragement given by Univer-
siti Teknologi Malaysia (UTM) and the Ministry of Educa-
tion Malaysia (MoEM) in making this research possible. This
academic work was supported by Fundamental Research Grant
Scheme (RJ130000.7816.4L088) initiated by MoEM.
References and Notes1. W. Baker, The ‘classroom flip’: Using web course management tools too
become the guide by the side, 11th International Conference on CollegeTeaching and Learning, Jacksonville, FL (2000).
2. B. Tucker, The flipped classroom, Education Next (2012), pp. 82–83.3. J. G. Jr, The Effects of a Flipped Classroom on Achievement and Student
Attitudes in Secondary Chemistry (2013).4. J. Chipps, The Effectiveness of Using Online Instructional Videos with Group
Problem Solving to Flip the Calculus Classroom (2013).5. M. J. Lage, G. J. Platt, M. Treglia, and J. Lage, J. Econ. Educ. 31, 30 (2000).6. C. F. Herreid and N. A. Schiller, J. Coll. Sci. Teach. 42 (2012).7. K. Lockwood, C. S. U. M. Bay, and R. Esselstein, The inverted classroom
and the CS curriculum, Proceeding of the 44th ACM Technical Symposiumon Computer Science Education, New York, USA (2013), pp. 113–118.
8. C. Demetry, Work in progress–An innovation merging classroom flip andteam-based learning, 40th IEEE Frontiers in Education Conference (FIE)(2010).
9. R. H. Rutherfoord and J. K. Rutherfoord, Flipping the classroom—Is it for you?14th annual ACM SIGITE Conference on Informtion Technology Education(2013), pp. 19–22.
10. N. Hamdan, P. McKnight, K. McKnight, and K. M. Arfstrom, Flip. Learn. Netw.(2013).
11. A. Butt, Bus. Educ. Accredit. 6, 33 (2014).12. J. Bergmann and A. Sams, Learn. Lead. with Technol. 36, 22 (2009).13. K. P. Fulton, Phi Delta Kappan J. Storage 94, 20 (2012).14. D. Siegle, Differentiating Instruction by Flipping the Classroom (2013), Vol. 37,
pp. 51–56.15. S. Flumerfelt and G. Green, Educ. Technol. Soc. 16, 356 (2013).16. N. K. Pang and K. T. Yap, The Flipped Classroom Experience, IEEE CSEE
and T 2014, Klagenfurt, Austria (2014), pp. 39–43.17. K. Peacock, Flipping Your Classroom?: A Ticket to Increased Classroom Col-
laboration?, Centre for Teaching and Learning University of Alberta (2013).18. D. N. Shimamoto, Implementing a flipped classroom?: An instructional mod-
ule, Technology, Colleges, and Community Worldwide Online Conference(2012).
19. B. B. Stone, Flip your classroom to increase active learning and studentengagement, 28th Annual Conference on Distance Teaching and Learning(2012), pp. 1–5.
20. A. Steed, Proquest 16, 9 (2012).21. J. F. Strayer, Learn. Environ. Res. 15, 171 (2012).22. N. Warter-Perez and J. Dong, Flipping the classroom?: How to embed inquiry
and design projects into a digital engineering lecture, Proceedings of the 2012ASEE PSW Section Conference (2012).
23. Z. Zhang, Construction of online course based on flipped classroom model(FCM) concept, 2nd International Conference on Information, Electronics andComputer (ICIEAC) 2014, September (2013), pp. 157–160.
24. D. Siegle, Gift. Child Today 37, 51 (2013).25. K. R. Clark, Examining the Effect of the Flipped Classroom Model of Instruc-
tion on Student Engagement and Performance in the Secondary Mathematicsclassroom: An Action Research Study (2013).
26. Y. Jiugen, X. Ruonan, and Z. Wenting, Essence of flipped classroom teachingmodel and influence on traditional teaching, IEEE Workshop on Electronic,Computer and Applications (2014), pp. 362–365.
27. J. A. Day and J. D. Foley, IEEE Trans. Educ. 49, 420 (2006).28. S. Kellogg, Developing online materials to facilitate an inverted class-
room approach, 39th ASEE/IEEE Frontiers in Education Conference (2009),pp. 1–6.
29. S. Zappe, R. Leicht, J. Messner, T. Litzinger, and H. W. Lee, Flipping the class-room to explore active learning in a large undergraduate course, AmericanSociety for Engineering Education (2009).
30. A. A. Rahman, N. Mohd Zaid, Z. Abdullah, H. Mohamed, and B. Aris,Emerging project based learning in flipped classroom, The 3rd InternationalConference of Information and Communication Technology (ICoICT) (2015),In revised.
Received: 30 November 2014. Accepted: 28 January 2015.
Data Store Architecture with Student Subject: Case
Study at Muhammadiyah University of Yogyakarta
Fajar Rianda∗, Asroni, and Ronald Adrian
Muhammadiyah University of Yogyakarta, Indonesia
Muhammadiyah University of Yogyakarta have an academic information system. This system is an Online Trans-action Processing (OLTP). System have data which related with student and can used for accreditation. But,system have not capability to give information quickly and accurately. Data warehouse is a system which havecapability to give information quickly and accurately. With this ability to easily get related information students.This research using Single Dimensional Data Store (DDS) architecture. The goal research is build student datamart on data warehouse of Muhammadiyah University of Yogyakarta using single DDS architecture and result-ing information quickly and accurately for supporting borang accreditation. Data mart has been built give easeto get information about student and the information can display in tabular.
Keywords: Data Warehouse, Data Mart, Borang Accreditation, Student, Single DDS.
1. INTRODUCTION1.1. Background
Every college have to accredited by Badan Akreditasi Nasional–
Perguruan Tinggi (BAN-PT). Accreditation is used to determine
eligibility and quality of college. BAN-PT will accrediting using
borang accreditation form.
Muhammadiyah University of Yogyakarta (UMY) as one of
the college who have to fill in the accreditation. UMY have an
academic information system. This system is an Online Trans-
action Processing (OLTP). System have data which related with
student and can used for accreditation. But, system have not
capability to give information quickly and accurately.
The solution for that case is data warehouse. Data warehouse
is a system which have capability to give information quickly
and accurately. With this ability to easily get related information
students.
1.2. Formulation of the Problem
Formulation of the problem based on the background above are
how to building data mart on data warehouse with single dimen-
sional data store architecture.
1.3. Goal
The goal of this research is to build student data mart at of
Muhammadiyah University of Yogyakarta using single DDS
∗Author to whom correspondence should be addressed.
architecture and resulting information quickly and accurately for
supporting borang accreditation.
2. THEORITICAL2.1. Literature Review
The Research related about data warehouse has been done several
times. Some research as a reference for this research is:
• Windarto has done research that entitle Pemanfaatan Data
Warehouse sebagai sarana Penunjang Penyusunan Borang
Akreditasi Standar 3 pada Fakultas Teknologi Informasi Univer-
sitas Budi Luhur.1 His conclusion research is using data ware-
house give ease to get information more effective and increase
data security.
• Mukhlis Febriady and Bayu Adhi Tama have done research
that entitle Rancang Bangun Data Warehouse untuk Menunjan
Evalusai Akademik di Fakultas.2 Their research resulting inte-
grated database and make report faster.
• Other research from Armadiyah Amborowati that entitle Per-
ancangan dan Pembuatan Data Warehouse pada Perpustakaan
STMIK AMIKOM Yogyakarta.3 Her research made load process
periodic automatically and help administrator.
2.2. Data Warehouse
Data Warehouse is a subject-oriented, integrated, time-variant,
nonupdatable collection of data used in support of management
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3392–3395, 2015
Fig. 1. Research procedure.
decision-making processes and business intelligence.4 The mean-
ing of each of the key terms in this definition follows:5
• Subject-oriented, a data warehouse is organized around the
key subjects.5 Major subjects may include students, patients, and
products.5
• Integrated, the source data from different source systems. The
source data usually inconsistent. The integrated source system
must be made consistent.6
• Time-Variant, Data in the warehouse is only accurate and valid
at some point in time or over some time interval.6
• Nonupdatable, Data in the data warehouse are loaded and
refreshed from operational system, but cannot be updated by end
users.5
2.3. Online Transaction Processing (OLTP)
Online Transaction Processing is a system whose main purpose
is to capture and store the business transactions.7 OLTP system
contain the data that will load into the data warehouse.7
2.4. Difference of OLTP and Data Warehouse
See Table I.
2.5. Data Mart
Data mart is a subset of a data warehouse that supports the
requirements of a particular department or business function.6
Fig. 2. Single DDS.
Fig. 3. Star schema.
Table I. Difference of OLTP and Data Warehouse.6
OLTP Data warehouse
Holds current data Hold historical dataData is dynamic Data is largely staticTransaction-driven Analysis-drivenApplication-oriented Subject-orientedSupports day-to-day decision Supports strategic decisions
2.6. Extract, Transform, and Load (ETL)
ETL is a system that has the capability to connect to the source
systems, read the data, transform data, and load it into a target
system.7
2.7. Dimensional Data Store
Dimensional Data Store is a database that stores the data ware-
house data in a different format than OLTP.7 DDS is a better
format to store data in the warehouse for the purpose of querying,
analyzing data and gives better query performance.7
2.8. Dimensional Modelling (Star Schema)
A Star schema is a simple database design in which dimensional
data are separated from fact or event data.5 A star schema consist
of two types of tables: fact tables and dimension tables. Star
schema is simpler than other schema, and making easier for ETL
process data into DDS.7
2.8.1. Single Dimensional Data Store (DDS)
Architecture
In Single DDS, consist of only two data stores: stage and DDS.
Single DDS architecture use two data store, which are:7
(a) A stage, is a place where you store the data you extracted
from the store system temporarily, before processing it further or
load into to other data store.
(b) Dimensional Data Store (DDS), is a user-facing data store,
where the data is arranged in dimensional format for purpose of
supporting analytical queries.7
Single DDS is simpler because the data from the stage is
loaded straight into the dimensional data store, without going to
any kind of normalized store first.7
Table II. Name changes from source store to stage.
After data have stored in stage data store, the next step is ETL
process. In second ETL process, data will stored from stage to
dimension data store. Beside store data, in this step will do data
quality. Data quality make sure data that stored in dimensional
data store are not dirty data or noise. Dirty data or noise can
include null data, duplicate data. In this step, name of tables
will change. For dbo.stage.faculty table and dbo.stage.department
table merged and became dbo.stage.dim_faculty_department.
Table III displaying tables name changes in DDS:
That tables are dimension table. In DDS, there are one fact
table. The table is dbo.fact_jumlah_ mahasiswa. Dimension table
and fact table made with star schema structure. Figure star
schema display in Figure 5.
As the result of database architecture that have made from
source system to dimensional data store is shown in the following
figure.
4.3. Testing
4.3.1. ETL Testing
In this testing will make sure that data in ETL process have
stored to data store. The results of testing display in Table IV.
4.3.2. Functional Testing
This testing make sure that data mart have capability to fulfill
requirement of borang accreditation document. Table V display
result of testing.
5. CONCLUSIONThe conclusion that can taken from this research is building
data mart give ease to get information about student, can used
by Muhammadiyah University of Yogyakarta to fill in borang
accreditation and the information can display in tabular.
References and Notes1. Windarto, Telematika MKOM 3 (2011).2. M. Febriady and B. A. Tama, Rancang bangun data warehouse untuk menun-
jang evaluasi akademik di fakultas, KNTIA, Palembang (2011).3. A. Amborowati, Perancangan dan pembuatan data warehouse pada per-
pustakaan stmik amikom yogyakarta, Seminar Nasional Aplikasi Sains DanTeknologi (2008).
4. W. H. Inmon, Building the Data Warehouse, Fourth edn., John Wiley and Sons(2005).
5. J. A. Hoffer, M. B. Prescott, and F. R. McFaden, Modern Database Manage-ment, Eightth edn., Pearson Education Inc. (2007).
6. T. M. Connolly and C. E. Begg, Database System A Practical Approach toDesign, Implementation, and Management, Fourth edn., edited by AddisonWeslet, United States of America (2005).
7. V. Rainardi, Building a Data Warehouse: With Example in SQL Server (2008).
Received: 17 December 2014. Accepted: 8 February 2015.
Nur Atika Jamuary1, Mohd Shoki Md Ariff1�∗, Hayati Jamaludin1, Khalid Ismail2,Nawawi Ishak3, and Mohd Sawal Abong3
1Faculty of Management, Universiti Teknologi Malaysia, 81300 Skudai, Malaysia2Universiti Pendidikan Sultan Idris, 35900 Tg. Malim, Malaysia
3Lembaga Tabung Haji Malaysia
The advancement of online channel has point out the importance of online brand trust in order to ensuresustainability of airline business. Online trust plays an important role in creating satisfied and expected outcomesbetween online buyers and airline companies, and the companies need to address factors contributing to theonline brand trust. The increase in international students’ intake in Malaysia provides huge opportunity foronline airline tickets service providers because they are flying to and from Malaysia on a continuous basis.However, studies on online brand trust are still limited in airline industry especially among international studentspursuing their study in Malaysia. A conceptual model of online brand trust featuring six factors—word of mouth,security/privacy, perceived risk, good online experience, brand reputation and perceived quality of information—was developed based on studies of brand trust to determine factors contributing to the online brand trust of aMalaysian-based airline company. Analysis of data was performed based on the 276 respondents, which werecollected using a convenience sampling method from international students in a public university of Malaysia.The result indicated that there are five factors contributing to the online brand trust in the airline industry—word of mouth, security/privacy, perceived risk, good online experience/brand reputation and perceived qualityof information. Good online experience/brand reputation appeared to have stronger effect on online brand trust.Theoretical and practical implications of this study were discussed in understanding online brand trust involvinginternational students and ways to enhance their trust towards the airline companies.
Information quality 0�341 0.046 0�309∗∗ 7.425 0�000 1.348
F 101.864 R 0.732 R2 0.808
Note: ∗p < 0�05; ∗∗p < 0�01.
3398
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3396–3399, 2015
related to the payment, available service, and et cetera is also a
contributing factor to online brand trust.
Word of mouth was found to have positive and significant
effect on online brand trust. This result is consistent with Ha,13
Kim and Song28 and Alam and Mohd Yasin17 who found pos-
itive and significant relationship between word of mouth and
online brand trust. Online buyers’ word mouth communication
is considered as important factor that would likely to influence
their online buying behavior in airline services. It shows that the
increasing of word of mouth that the brand has, the higher the
level of trust that customer has toward that brand.
Security was positively and significantly affected online brand
trust, which is consistent with the works of Ha,13 Srinivasan16
and Alam and Mohd Yasin.17 When privacy rules are clearly
defined in the airline company’s website and customers do not
have any worries on providing of their personal information, it
indicate that they are more likely to trust the airline brand. This
indicated that the more customers feel secure about online brand
and have confidence in providing their personal information, the
more they trust that online brand.
Furthermore, perceived risk was also found to have positive
effect on online brand trust. However, this result contradicts with
those of previous studies done by Alam and Mohd Yasin17 and
Mohammadian and Ghanbar.18 Previous studies have found that
perceived risk has no significant influence on online brand trust.
This contradicting result might be based on the fact that many of
the international students’ perception of risk regarding of buying
online ticket from the airline company’s website may not have
impact on their trust of that online brand.
The results had shown that good online experience/brand rep-
utation had positive effects on online brand trust. This result
was also observed in the previous research carried out by Ha,13
Ruparelia et al.,6 Alam and Mohd Yasin17 and Mohammadian
and Ghanbar.18 Online buyers are seemed to have trusted on the
airline brand and they are more likely to repurchase from the air-
line website because they have a pleasant experience dealing with
the online purchase of air ticket. Therefore, good online experi-
ence/brand reputation is a factor contributing to higher trust in
online brand.
Previous studies conclude that the higher the quality of infor-
mation that online brand’s website offers, the higher the level of
brand trust the consumer gained.6�13�17�18 This study supports this
view in which perceived quality of information positively affect
online brand trust. It calls online companies to focus more on
providing information that can satisfy and fulfil their online cus-
tomers’ needs instead of providing unnecessary and unorganized
information.13
6. CONCLUSIONSThis study concluded that the contributing factors to online brand
trust in airline industry are word of mouth, security/privacy, per-
ceived risk, online experience/brand reputation and perceived
quality of information. These factors are very much related to
the perception of online buyers, and may not be sufficient to
determine brand trust in online setting. In the purchase of air-
line ticket, buyers may encounter difficulties related to the func-
tionality of the website. Therefore, factor of web design and
navigation6�18 should be considered in examining online brand
trust. Advertising and testimonial6 could also be used in the con-
ceptual model of online brand trust to comprehensively capture
all possible factors that influence online buyers to develop their
trust in online setting.
In this study, international students who are pursuing their
studies in Malaysia had expressed their views on factors con-
tributing to the online brand trust. Further study should be carried
out to generalize contributing factors of online brand trust in air-
line industry by extending participation of various segments of
online air ticket.
References and Notes1. D. H. McKnight, V. Choudhury, and C. Kacmar, Information System Research
13, 334 (2002).2. D. Buhalis and R. Law, Progress in information technology and tourism man-
agement: 20 years on and 10 years after the Internet—The state of eTourismresearch, Tourism Management (2008), Vol. 29, pp. 609–623.
3. E. Bigne, B. Hernandez, and L. Andreu, Journal of Air Transport Management16, 346 (2010).
4. C. W. Wang, Airline ticket E-Reservation: Adoption among Malaysians, Masterthesis, USM (2011).
5. Malaysian Communication and Multimedia Commission (2008),www.skmm.gov.
6. N. Ruparelia, L. White, and K. Hughes, Journal of Product and BrandManagement 19, 250 (2010).
7. S. L. Jarvenpaa, N. Tractinsky, and M. Vitale, Information Technology andManagement 1, 45 (2000).
8. D. Louis and C. Lombart, Journal of Product and Brand Management 19, 114(2010).
9. E. Delgado-Ballester, and J. L. Munuera-Aleman, Journal of Product andBrand Management 14, 187 (2005).
10. M. Horppu, O. Kuivalainen, A. Tarkianen, and H. K. Ellonen, Journal of Prod-uct and Brand Management 17, 403 (2008).
11. A. Mukherjee and P. Nath, European Journal of Marketing 41, 1173 (2007).12. J. J. Zboja and C. M. Voorhees, Journal of Services Marketing 20, 381 (2006).13. H. Y. Ha, Journal of Product and Brand Management 13, 329 (2004).14. R. M. Morgan and S. D. Hunt, Journal of Marketing 58, 20 (1994).15. R. E. Spekman, Strategic Supplier Selection: Understanding Umg Term Buyer
Relationships, Business Horizons, (July/August) (1988), pp. 75–81.16. Srinivasan, Information Management and Computer Security 12, 66 (2004).17. S. S. Alam and N. M. Yasin, Journal of Theoretical and Applied Electronic
Commerce Research 5 (2010).18. M. Mohammadian and M. Ghanbar, Switzerland Research Park Journal 103
(2014).19. N. Ozguven, Chinese Business Review 10, 990 (2011).20. J. C. Roca, J. J. Garcia, and J. J. Vega, Information Management and Com-
puter Security 17, 96 (2009).21. W. D. Salisbury, R. A. Pearson, A. W. Pearson, and D. W. Miller, Industrial
Management and Data Systems 101, 165 (2001).22. C. Flavian and M. Guinaliu, Industrial Management and Data Systems
106, 601 (2006).23. H. H. Chang and S. W. Chen, Online Information Review 32, 818 (2008).24. W. Huang, H. Schrank, and A. J. Dubinsky, Journal of Consumer Behaviour
4, 40 (2004).25. M. Laroche, M. V. Nepomuceno, and R. Marie-Odile, Journal of Consumer
Marketing 27, 197 (2010).26. K. H. Hahn and J. Kim, International Journal of Retail and Distribution
Management 37, 126 (2009).27. L. V. Casalo, C. Flavian, and M. Guinaliu, The International Journal of Bank
Marketing 26, 399 (2008).28. H. Kim and J. Song, Journal of Research in Interactive Marketing 4, 376
(2010).29. C. Kohli, L. Leuthesser, and S. Rajneesh, Got slogan? Guidelines for creating
effective slogans, Business Horizons (2007), pp. 415–422.30. L. C. Harris and M. H. Goode, Journal of Services Marketing 24, 230 (2010).31. Y. H. Chen and S. Barnes, Industrial Management and Data Systems 107, 21
(2007).32. J. F. Hair, Jr., B. Babin, H. A. Money, and S. Philip, Essentials of Business
Research Methods, Wiley: Malloy Inc., USA (2003).
Received: 17 December 2014. Accepted: 8 February 2015.
Li Yuan Hui1, Mohd Shoki Md Ariff1�∗, Norhayati Zakuan1, Norzaidahwati Zaidin1,Khalid Ismail2, and Nawawi Ishak3
1Faculty of Management, Universiti Teknologi Malaysia, 81300 Skudai, Malaysia2Universiti Pendidikan Sultan Idris, 35900 Tg. Malim, Malaysia
3Lembaga Tabung Haji Malaysia
The online shopping development in China has attracted many researchers to examine technology and humanbehavior factors affecting consumer online attitude in e-commerce setting. In this study, the conceptualizationof technology and trust factors to determine attitude of online shoppers was based on Technology AcceptanceModel (TAM) and Theory of Planned Behavior (TPB). Perceived Ease of Use (PE) and Perceived Useful-ness (PU) of technology factor and trust in website and perceived risk were constructed to determine attitude ofonline shoppers of the biggest network for online shopping in China. A web-based questionnaire was employedto collect 260 feedbacks from online shoppers of the website. Both constructs of technology factor were foundto have significant and positive effect on attitude of online shoppers. The online shopper’s attitude was sig-nificantly and positively influenced by their trust in the website. The effect of perceived risk of trust factor onconsumer attitude was insignificant. This study revealed the importance of integrating technology and trust fac-tors in understanding attitude of online shoppers. This finding calls for online retailers and e-marketers to focuson both technology and human dimensions in their marketing effort to facilitate more people to shop online.
Keywords: TAM, Theory of Planned Behavior, Attitude, Trust, Online Purchase Intention.
1. INTRODUCTIONNowadays, online shopping has become one of the essential
characteristics in the Internet era. Online shopping has become
the third most popular Internet activity after e-mail and web
browsing,1 and it is even more popular than seeking out enter-
tainment, information and news. For online consumers, online
shopping brings a great number and wide range of merchandises
whilst it offers a huge market and numerous business opportu-
nities for online retailers. Online shopping in China is devel-
oping tremendously, in which the online shopping had risen to
142 million, and the volume of the online transaction rose to
RMB523.1 billion in 2010.2 As the world’s largest online pop-
ulation, China is set to become the world’s largest online retail-
ing market.3 Further, the sales volumes online retail market in
China, based on the forecasts by The Boston Consulting Group,
is expected to reach 364 billion US dollars by 2015.4
The online shopping development in China has attracted many
researchers to examine technology and human behavior fac-
tors affecting consumer attitude. Online shopping involves new
technologies in order for online shoppers to browse, search,
compare, and finally making a purchase decision, thus under-
standing how consumers’ perceived the websites technology
∗Author to whom correspondence should be addressed.
and human behavior in online environment are important to
researchers and online retailers.4 In understanding attitude of
online shoppers, the trust and technology factors has been well
studied in on-line shopping and showed that understanding both
the technology and trust factors is important in determining con-
sumer behavioral intention.5 However, most previous studies in
China assessing trust and technology factors and their influence
on consumer attitude addressed the issues separately. This fact
poses some difficulties for e-marketers and online retailers to
incorporate issue of technology and human behavior or trust fac-
tor in designing marketing strategy and program. Thus, integrat-
ing both technology and trust factors is needed to better explain
consumers attitude in China’s online shopping context.
Taoboa6 is one of the shopping website in China with 76.5%
market share in 2009, making it the most preferences website
among online shoppers.2 Since its launching in 2003, Taoboa is
the largest Internet retail and trading website in China, offer-
ing consumers with wide range of products and services.4 It is
one of the world’s largest electronic marketplaces, with over
370 million registered users and annual transaction volume of
almost $15 billion in 2008, equaling approximately one percent
of China’s total retail trade.4 Thus, revealing technology and trust
factors in understanding attitude of online shoppers at the Taoboa
transaction. For website designers, this fact means designing
consumer-friendly website is crucial;
(ii) in in-store purchase, consumers will visit the outlets,
search, check, touch and finally commit in a purchase of a
product. This process may take minutes to hours depending
on type of product purchased. In online buying, if the online
process take less time compared to that of in in-store pur-
chase, consumers may perceived the online purchase to be
faster, time saving, quick and very useful. This point must be
well addressed in marketing promotional mix to attract online
buyers continuing the online purchase.
• The negative effect of perceived risk on consumer attitude
towards online shopping is surprisingly insignificant. This finding
is inconsistent with the previous researches of Van der Heijden
et al.,9 O’Cass28 and Shih.29 This study also addressed trust in
website and this part of trust was positively influenced the atti-
tude of online shoppers. Trust includes consumers’ belief that
Taoboa website is capable and proficient in its business. In con-
trast, some aspects of perceived risk addressed potential loss and
gain and the likelihood of making a good bargain through online
purchase, which is difficult to assess by respondents, thus may
contribute to the insignificant effect of perceived risk on the
attitude.
6. CONCLUSIONSThe issue of how technology and trust factors affect attitude of
online shoppers provide additional insight in the study of attitude
of online shoppers, particularly in China online shopping. These
two factors are proven to have significant effect on the attitude,
and the variation in the attitude of online shoppers is very much
influenced by these two factors. Perceived risk of trust factor is
insignificant, and this requires researchers to further validate this
finding. Sampling procedure pose another limitation of this study
because background of the respondents, particularly frequency
and experience of online shoppers are not considered in the sam-
pling frame and strategy. Previous studies highlighted that these
two aspects influenced perception of online shoppers. Three vari-
ables in this study are related to how consumer perceived ease
of use and usefulness as well as perceived risk, therefore, the
inclusion of these two points would be interesting to research.
References and Notes1. N. Li and P. Zhang, Consumer online shopping attitude and behaviour, Eight
American Conference on Information System (2002), pp. 508–517.2. G. Jun and N. Jaafar, International Journal of Business and Social Science
2, 122 (2011).3. K. Michael, China’s Online Retail Market to Triple to $364bB by 2015, Says
Research Firm, IDG News, April 12 (2012).4. D. Zhao, The Chinese consumer shopping behavior on taobao, unpublished
thesis, Uppsala Universitet (2012).5. T. L. Wang and Y. F. Tseng, International Journal of Digital Society 2, 433
(2011).6. www.taoboa.com.7. M. S. Md Ariff, N. Zakuan, Y. S. Min, K. A. Rahim, and K. Ismail, International
Journal of Information Processing and Management (IJIPM) 4, 48 (2013).8. F. D. Davis, MIS Quarterly 13, 319 (1989).9. H. van der Heijden, T. Verhagen, and M. Creemers, European Journal of IS
12, 41 (2003).10. I. Ajzen, Organisational Behaviour and Human Decision Processes 50, 179
(1991).11. M. S. Ariff, S. M. Yeow, and N. Zakuan, Adv. Sci. Lett. 20, 268 (2013).12. T. Ramayah and J. Ignatius, Impact of perceived usefulness, perceived ease
of use and perceived enjoyment on intention to shop online, 1 (2005).
3403
R ES E A R CH AR T I C L E Adv. Sci. Lett. 21, 3400–3404, 2015
13. T.-L. Wang and Y. F. Tseng, International Journal of Digital Society (IJDS)2, 433 (2011).
14. L. Zheng, M. Favier, P. Huang, and F. Coat, Journal of Electronic CommerceResearch 13, 255 (2012).
15. F. D. Davis, R. P. Bagozzi, et al., Management Science 35, 983 (1989).16. F. D. Davis, Bagozzi, and Warshaw, Journal of Applied Social Psychology
22, 1111 (1992).17. M. S. Feathermana and P. A. Pavloub, Int. J. Human-Computer Studies
59, 451 (2003).18. V. Venkatesh and F. D. Davis, Management Science 46, 186 (2000).19. D. Gefen, E. Karahanna, and D. Straub, 27, 51 MIS Q (2003).20. K. C. Ling, L. T. Chai, and T. H. Piew, International Business Research 3, 63
(2010).21. McKnight, Choudhury, and Kacmar, Information Systems Research 13, 334
(2002).22. S. L. Jarvenpaa and N. Tractinsky, Information Technology and Management
1, 45 (2000).
23. G. Hofstede and M. H. Bond, Journal of Cross-Cultural Psychology 15, 417(1984).
24. K. S. Lee and S. J. Tan, Journal of Business Research 56, 877 (2003).25. M. S. Md Arif, M. Sylvester, K. A. Rahim, and N. Zakuan, Adv. Sci. Lett.
20, 2319 (2014).26. V. W. Mitchell, European Journal of Marketing 33, 163 (1999).27. M. S. Md Arif, M. Sylvester, N. Zakuan, K. Ismail, and K. M. Ali, IOP
Conference Series: Materials Science and Engineering 58, 1 (2014).28. A. O’Cass and T. Fenech, Journal of Retailing and Consumer Services 10,
81 (2003).29. H. P. Shih, Information and Mgt. 41, 351 (2004).30. P. G. Zimbardo and J. N. Boyd, J. Personality Soc. Psychol. 77, 1271 (1999).31. Y. B. Chiu, C. P. Lin, and L. L. Tang, International Journal of Service Industry
Management 16 (2005).32. I. Ajzen, Annual Review of Psychology 52, 27 (2001).33. J. F. Hair, R. E. Anderson, et al., Multivariate Data Analysis. Prentice-Hall
(1998).
Received: 17 December 2014. Accepted: 8 February 2015.
Daniel Hartono Sutanto∗ and Mohd. Khanapi Abd. Ghani
Biomedical Computing and Engineering Technologies (BIOCORE) Applied Research Group;Faculty of Information and Communication Technology, Universiti Teknikal Malaysia Melaka, Melaka, 76100, Malaysia
Non-communicable disease (NCDs) is the most epidemic disease and high mortality rate in worldwide likelydiabetes mellitus, cardiovascular diseases and cancers. NCDs prediction model have problems such as redun-dant data, missing data, imbalance dataset and irrelevant attribute. In data mining, feature selection can handleirrelevant attribute. This paper considers finding the optimal feature selection for NCDs prediction model. Wecomprise 18 feature selection, 4 classification algorithms (Naïve Bayes, Support Vector Machine, Neural Net-work, and Decision Tree) and used 6 NCDs datasets. The result shows that optimally performed feature selectionfor NCDs prediction are weight by SVM, W-Uncertainty, W-Chi, and CBWA.
Let r ij be the rank of the j-th of C algorithms on the i-th
of D datasets. The Friedman test has an aim to compare the
mean ranks of algorithm Rj = �1/D�∑D
i−1 rij . Under the null-
hypothesis, which says that all the algorithms are equivalent and
so their ranks Rj should be clean. The statistic of Friedman is
calculated as follows, and disseminated according to x2F with C
−1 degrees of freedom, when variable D and C are big enough.
x2F = 12D
C�C+1�
[ D∑j
R2j −
C�C+1�2
4
](23)
If the null-hypothesis is rejected, it can go along with a post-hoc
test. When all classifiers are compared to each other, the Nemenyi
test should be applied. Two classifiers have significantly different
performance if the corresponding average ranks differ by at least
the critical difference, shown by
CD = qa
√C�C+1�
D(24)
where critical values qa are based on the studentized range
statistic.
Table XII. Significant differences of nemenyi post hoc test.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
1 N N N N N N N N N Y N N N N N N N N2 N N N N N N N N N N N N N N N N N N3 N N N N N N N N N N N N N N N N N N4 N N N N N N N N N N N N N N N N N N5 N N N N N N N N N Y N N N N N N N N6 N N N N N N N N N N N N N N N N N N7 N N N N N N N N N Y N N N N N N N N8 N N N N N N N N N N N N N N N N N N9 N N N N N N N N N N N N N N N N N N10 Y N N N Y N Y N N N N N N N N N N N11 N N N N N N N N N N N N N N N N N N12 N N N N N N N N N N N N N N N N N N13 N N N N N N N N N N N N N N N N N N14 N N N N N N N N N N N N N N N N N N15 N N N N N N N N N N N N N N N N N N16 N N N N N N N N N N N N N N N N N N17 N N N N N N N N N N N N N N N N N N18 N N N N N N N N N N N N N N N N N N
3.8. Experimental Infrastructure
In this research, the experiment equipped with infrastructure con-
sists RapidMiner Toolkit is an open-source system consisting
of a bit of data mining algorithms to automatically examine a
large data collection and extract useful knowledge. The XLSTAT
statistical analysis add-in offers a wide variety of purposes to
enhance the analytical capacities of Excel, producing it the ideal
puppet for your everyday data analysis and statistics prerequi-
sites. The hardware used CPU: HP Z420 Workstation, Proces-
Bayes, Support Vector Machine, Neural Network, and Decision
Tree) and used 6 NCDs datasets. The result shows that opti-
mally performed feature selection for NCDs prediction is weight
W-SVM, W-Uncertainty, W-Chi, and CBWA.
Acknowledgment: This work was supported with a
grant from LPDP Minister of Finance of Indonesia No.
Kep56/LPDP/2014.
References and Notes1. WHO, Global Status Report on Noncommunicable Diseases (2010).2. M. K. A. Ghani, R. K. Bali, R. N. G. Naguib, I. M. Marshall, and N. S.
Wickramasinghe, Int. J. Healthc. Technol. Manag. 11, 113 (2010).3. D. H. Sutanto, N. S. Herman, and M. K. A. Ghani, Adv. Sci. Lett. 20, 1740
(2014).4. I. Guyon, J. Mach. Learn. Res. 3, 1157 (2003).5. V. Bolón-Canedo, N. Sánchez-Maroño, and A. Alonso-Betanzos, Knowl. Inf.
Syst. 34, 483 (2013).6. N. Yilmaz, O. Inan, and M. S. Uzer, J. Med. Syst. 38, 48 (2014).7. R. S. Wahono, N. S. Herman, and S. Ahmad, Adv. Sci. Lett. 20, 1945
(2014).8. Y.-W. Chang and C.-J. Lin, Causation Predict. Chall. 53 (2008).9. I. Kononenko, Machine Learning: ECML-94 784, 171 (1994).
10. M. A. Hall and L. A. Smith, Comput. Sci. 98, 181 (1998).11. M. A. Hall, Methodology 21i195-i20, 1 (1999).12. L. Yu and H. Liu, Efficient Feature Selection via Analysis of Relevance and
Redundancy 5, 1205 (2004).13. C. Ding and H. Peng, J. Bioinform. Comput. Biol. 3, 185 (2005).14. L. Yu and H. Liu, Int. Conf. Mach. Learn. 1 (2003).15. R. Tibshirani, T. Hastie, B. Narasimhan, and G. Chu, Proc. Natl. Acad. Sci.
U. S. A. 99, 6567 (2002).16. I. Guyon, J. Weston, and S. Barnhill, Mach. Learn. 46, 389 (2002).17. P. Luukka, Expert Syst. Appl. 38, 4600 (2011).18. S. Belciug and F. Gorunescu, J. Biomed. Inform. 52, 329 (2014).19. L.-Y. Chuang, C.-H. Yang, K.-C. Wu, and C.-H. Yang, Comput. Biol. Med.
41, 228 (2011).20. D.-C. Li, C.-W. Liu, and S. C. Hu, Artif. Intell. Med. 52, 45 (2011).21. Y. J. Fan and W. A. Chaovalitwongse, Ann. Oper. Res. 174, 169 (2010).22. N. A. Mat Isa and W. M. F. W. Mamat, Appl. Soft Comput. 11, 1457 (2011).23. M. A. Chikh, M. Saidi, and N. Settouti, J. Med. Syst. 36, 2721 (2012).24. F. Beloufa and M. A Chikh, Comput. Methods Programs Biomed. 112, 92
(2013).25. J. Zhu, Q. Xie, and K. Zheng, Inf. Sci. (Ny) 292, 1 (2015).26. Data Mining Pratical Machine Learning Tools and Techniques, Ian H. Witten
Eibe Frank Mark A. Hall, 3rd edn. (2011).27. S. Lessmann, B. Baesens, C. Mues, and S. Pietsch, IEEE Trans. Softw. Eng.
34, 485 (2008).28. V. Van Belle and P. Lisboa, Artif. Intell. Med. 60, 53 (2014).29. J. Demšar and J. Demšar, J. Mach. Learn. Res. 7, 1 (2006).
Received: 17 December 2014. Accepted: 8 February 2015.
Yap Soon Jing1, Norzaidahwati Zaidin1, Mohd Shoki Md. Ariff1�∗, Norhayati Zakuan1,Khalid Ismail2, and Nawawi Ishak3
1Faculty of Management, Universiti Teknologi Malaysia, 81300 Skudai, Malaysia2Universiti Pendidikan Sultan Idris, 35900 Tg. Malim, Malaysia
3Lembaga Tabung Haji Malaysia
Online shopping is now a trend in e-business across the world, and the website is the main platform for thee-business. Thus, website quality is one of the crucial factors to determine the success of online business.This study aims to determine the website quality and how it fosters the attitude of online shopping amongyoung generation (Y-Gen). Based on the review of the past researches, WebQual was identified as relevantand the framework of this research was developed based on it. Altogether, 39 items of four website qualitywere treated as the key constructs (Usefulness (U), Ease of Use (EoU), Entertainment (E) and ComplimentaryRelationship (CR)). Stratified sampling method and questionnaire were adopted as the sampling procedure anddata collection purposes. The result showed that the four dimensions of website quality were positively andsignificantly affect attitude of Y-Gen towards the online shopping and EoU appeared to be the most dominantpredictor of attitude of online shoppers. Theoretical and managerial implications of the study are presented inthis article.
R E S E A R CH A R T I C L EAdv. Sci. Lett. 21, 3418–3421, 2015
entertainment, complimentary relationship and ease of use are
very crucial in determining attitude of online buyers.
Among the four construct of website quality, ease of use was
found to have stronger effect on attitude of Gen Y in the online
shopping setting. This finding synchronises with what had been
addressed by Yulihasri and Daud28 and a research by Lim and
Ting.27 The underlying factors are that online buyers tend to have
positive attitude towards the online shopping if they feel that the
online shopping website is easy for them to search a product,
place an order and deal with the online sellers. This phenomenon
is universal as it was evidenced in the many previous studies as
mentioned earlier. Thus, regardless of generation of users, how
online buyers perceived easiness of accessing and using the web-
site will contribute to the positive attitude towards the online
shopping. According to WebQual, Ease of Use is measure web-
site the degree of easy to read, operate and understand. The cru-
cial factors of information quality are the understandability and
format of information which is the way information is presented
to consumers.34 Thus, the content of website easy to read and
understand could result consumer perceived that the information
provided by the website to be high quality. This has suggested
that online shopping company could emphasize Ease of Use to
improve the quality of the website.
Childers et al.24 stated perceived usefulness is the stronger
predictors of consumer attitude towards online shopping. This
study supported the idea in which perceived usefulness is the key
construct that help users perform the task efficiently in term of
time and communication, consistent with the findings of Gefen25
and Bisdee.35 It shows that Gen Y will prefer to use the web-
site if the website could enable them to accomplish their task.25
Similar finding is observed among the Gen Y, therefore, it can
be concluded that this phenomenon is universal among online
users.
6. CONCLUSIONSThe result has shown that the website quality positively affect
Gen Y attitude towards online shopping. This study concludes
that it supported major theories addressing website quality and
attitude of online buyers like TAM and TRA. It is suggested that
future study to extend other groups of Gen Y prior to gener-
alizing the universal issue of website quality and online buyers
attitude. Further, this conclusion is made by comparing this find-
ing with previous researches involving general users. Therefore,
comparing among generation of online buyers is recommended.
References and Notes1. K. C. Williams and R. A. Page, Journal of Behavioural Studies in Business 1
(2010).2. Consumer Behavior Report, Online Purchasing Trends by Generation from
http://www. PriceGrabber.com (2008), p. 1.
3. J. M. Wolburg and J. Pokrywczynski, Journal of Advertising Behaviour 4, 33(2001).
4. S. Jones, The Internet goes to college: How students are living in the futurewith today’s technology. Pew Internet and American Life Project, September(2002), Available at: www.pewinternet.org.
5. N. Delafrooz, L. Paim, and A. Khatibi, Journal of American Science 6, 137(2010).
6. E. Wong, College back to school spending up 13%. Brandweek July (2010),Vol. 7.
7. D. Lester, A. M. Forman, and D. Lloyd, Services Marketing Quarterly 27, 123(2006).
8. Economic Planning Unit Malaysia from (2010). http://www.epu.gov.my/html/themes/epu/epu/images/common/pdf/ec Retrieved December 2013,
9. M. F. Sabri, M. MacDonald, J. Masud, L. Paim, T. Hira, and M. Othman,Consumer Interests Annual 54, 166 (2008).
10. S. J. Barns and R. T. Vidgen, Journal of Electronic Commerce Research3, 114 (2002).
11. Li, Hairong, T. Daugherty, and F. Biocca, Journal of Interactive Marketing15, 13 (2001).
12. B. Bai, R. Law, and I. Wen, International Journal of Hospitality Management27, 391 (2008).
13. D. Cyr, Journal of Management Information System 20, 33 (2008).14. Y. Boonghee and N. Donthu, Quarterly Journal of Electronic Commerce 2, 31
(2001).15. V. Zeithaml, Managing Service Quality 12, 135 (2002).16. A. M. Aladwania and P. C. Palvia, Information and Management 39, 467
(2002).17. J. Santos, Management Service Quality 13, 233 (2003).18. E. T. Loiacono, R. T. Watson, and D. L. Goodhue, WebQual: A web
site quality instrument, Working Paper 2000-126-0, University of Georgia(2002).
19. H. H. Bauer, T. Falk, and M. Hammerschmidt, Journal of Business Research59, 866 (2006).
20. M. Wolfinbarger and M. C. Glly, Comq: Dimensionalizing, Measuring and Pre-dicting Quality of the E-tailing Experience, Working Paper No.02–100, Mar-keting Science Institute (2002).
21. F. D. Davis, MIS Quarterly 13, 319 (1989).22. J. C. Lam and A. Parasuraman, Journal of Interactive Marketing 22, 19
(2008).23. J. W. Moon and Y. G. Kim, Extending the TAM for a World-Wide-Web cntext,
Inofrmation and Management (2001), Vol. 38, pp. 217–230.24. T. L. Childers, L. C. Carr, J. Peck, and S. Carson, Journal of Retailing 77, 511
(2001).25. D. Gefen, D. W. Straub, and M. C. Boudreau, Communications of the Associ-
ation for Information System 4, 2 (2000).26. Z. Selamat, N. Jaffar, and B. H. Ong, European Journal of Economics,
Finance and Administrative Sciences 1, 143 (2009).27. W. M. Lim and D. H. Ting, Journal of Modern Applied Science 6, 49
(2012).28. I. A. Yuslihasri and A. K. Daud, International Journal of Marketing Studies
3, 128 (2011).29. T. P. Monsuwe, G. C. B. Dellaert, and K. D. Ruyter, International Journal of
Service Industry Management 15, 102 (2004).30. B. J. Corbitt, T. Thanasankit, and H. Yi, Electronic Commerce Research and
Applications 2, 203 (2003).31. Heijden, V. D. Hans, T. Verhagen, and M. Creemers, European Journal of
Information Systems 2, 41 (2003).32. www.utm.my.33. R. Krejcie and D. Morgan, Determining Sample Size for Research Activities:
Education and Physiological Measurement, Upper Saddle River, N. J, PrenticeHall (1970).
34. Y. Salaün and K. Flores, International Journal of Information Management21, 21 (2001).
35. D. Bisdee, Consumer Attitudes Review, Office of Fair Trading, June (2007),pp. 1–147.
Received: 7 January 2015. Accepted: 20 February 2015.
Characteristics of Trustees and Trustors Affecting
Consumer Trust in Online Purchasing
Nur Shafiqah Ghazali1, Mohd Shoki Md. Ariff1�∗, Khalid Ismail2, Abdul Halim Ali2,Amir Hasan Dawi2, and Nawawi Ishak3
1Faculty of Management, Universiti Teknologi Malaysia, 81300 Skudai, Malaysia2Universiti Pendidikan Sultan Idris, 35900 Tg. Malim, Malaysia
3Lembaga Tabung Haji Malaysia
In e-commerce, trust is a critical success factor influencing the acceptance of online purchasing. Online shop-ping involves both online sellers and buyers, and their characteristics are important in determining the success ofinternet-based businesses. Purchase intention behavior of consumers is greatly affected by trust and thereforeexamining characteristics of trustors and trustees influencing consumer trust is needed. In this study, mea-surements of consumer trust include both a trusting party (trustor) and a party to be trusted (trustee). Thecharacteristics of trustees are measured based on perceive reputation; perceive size, and system assuranceof online sellers, while characteristic of trustors (online consumers) is determined by propensity to trust. Self-administered questionnaire was employed to collect 250 (83% response rate) completed questionnaire frombusiness students of a public university in Malaysia. In online purchasing, perceived reputation and systemassurance of trustees and propensity to trust of trustors positively and significantly affected consumer trust. Per-ceives reputation exerted higher effect on consumer trust, highlighting how consumer perceived credibility andreputation of online sellers is vital in forming trust-related behavior in online shopping. Perceived size of onlinesellers business was found to have a negative and insignificant effect on consumer trust. Purchase intention ofconsumers was significantly and positively affected by their trust towards the online purchasing. Theoretical andpractical implications of the study were discussed.
Keywords: Consumer Trust, Commitment-Trust Theory, Theory of Planned Behavior, Purchase Intention.
1. INTRODUCTIONIn online purchasing, online buyers will be attracted to buy
products using electronic service channels if they trust the seller
or the website. Trust plays an important role in online purchasing,
and consumers’ purchase intention was positively influenced by
their trust towards the online sellers.1 According to Teo and Liu,2
online purchasing has limitation such as the physical separation
between consumer and online seller, and between consumer and
product or service. This limitation created many problems or
uncertainties in online purchasing, and one of the main prob-
lems is lack of consumer trust.3 Even though most people are
choosing to shop online, but some of them did not do it because
of they didn’t trust the online sellers. Online buyers who shop
online concern more on trust because they can’t touch and feel
the products, but they had to pay for the products before they
are sent to them. Some of online buyers did not trust online
shopping because they can’t touch and feel the product, issues
of security of the website and lack of trust on the characteris-
tics of online seller and their online shopping website.4 In order
to reduce this barrier, online sellers must develop a trustworthy
∗Author to whom correspondence should be addressed.
relationship to gain customer loyalty and long term relationship
with online buyers.
Previous researches addressing consumer trust in online setting
assessed this issue from technology perspective, such as per-
ceived ease of use and perceive usefulness,5–7 and perceive
credibility.7 Others view it from human behavior perspective, for
example perceived risks and trust in website.5�8 However, the
consumers’ trust in online purchasing also develops based on
how they judge characteristics of online sellers and their propen-
sity to trust.2 Chen and Dhillon9 suggest that consumers will
trust in online purchasing based on situation or belief that the
online vendor is reliable, i.e., how they perceive characteristics
of online sellers. The antecedent of consumer trust in the con-
text of online purchasing is based on characteristics of trustees
such as perceive reputation, perceive size, system assurance; and
characteristic of trustors which is propensity to trust.2 These char-
acteristics are related with the trust that effect customer decision
where the consumers’ attitudes toward online seller will affect
their willingness to buy. Both characteristics are important deter-
minant of trust in online purchasing. Therefore, examining char-
acteristics of trustees and trustors to determine consumer trust
human behavior factors in understanding consumer trust. This
finding is consistent with many previous studies.2�9�11�22
Perceive reputation of online sellers significantly contribute to
consumer trust, which is consistent with previous studies.2�15�16
It shows that when online buyers perceived online sellers as
being honest, fair and consumer oriented, the trust related behav-
ior would be developed. From the system assurance perspective,
this research highlights that consumer still believe that security
of transaction system is important in online purchasing.2�13�15–17
Therefore, it is important for online sellers to ensure that the
online transaction systems must be stable, reliable, dependable
and secure. System approach deals with the extent to which
online buyers’ belief the online transaction of online sellers is
stable, reliable, dependable and secure,2�7 which is related to the
technology or websites quality of online sellers. Perceived rep-
utation relates to reputation of online sellers in term of having
good reputation and being honest, consumer oriented and faired
to online buyers, which is concerned to individual buyers. Inte-
grating these two perspectives provides a better understanding
of consumer trust in online setting. It implies that, in stimulat-
ing online trust, focusing on system assurance of technology or
website factor only is insufficient.
The propensity to trust of characteristic of trustors has positive
effect on consumer trust, which is consistent with the previ-
ous research findings.2�9 This study shows that, in online shop-
ping, individual’s general perceptions of whether or not they can
believe others play an important role in developing trust related
behavior. Thus, creating situation in which online buyers can eas-
ily develop trust towards the online sellers should be emphasized.
In addition, positive relationship exists between consumer
trusts with purchase intention. This finding synchronises with
what had been addressed by Kim et al.11 and Chen and Barnes.16
This shows that when consumers trust the online sellers, they will
have intention to make a purchase. The trust-purchase intention
relationship is observed in many studies worldwide, and therefore
it can be concluded that consumers’ trust precedes their inten-
tion to commit in online purchasing. E-marketers should develop
their marketing strategies based on the trust-intention relation-
ship, such as highlighting firms’ reputation in e-commerce setting
and build confident in the usage companies’ websites to stimulate
online purchasing.
6. CONCLUSIONSThis study enriches literature on trust in online shopping by
examining both characteristics of online sellers and online buy-
ers to determine consumer trust, as well as the effect of trust
on purchase intention. The finding suggests that addressing both
characteristics of trustees and trustors is important in under-
standing trust related behavior of consumer in online shopping.
For marketers, effective trust-building mechanisms based on the
characteristics of trustees and trustors should be embedded in
their business strategies.
This study addresses consumer trust from the business stu-
dents’ point of view, in which both characteristics of trustees
and trustors are important determinants of consumer trust in
online shopping. Business students are well exposed to both
e-commerce and internet-based information system; therefore,
their view on the role of trustees and trustors characteristics
in developing consumer trust should not be neglected. Future
research comparing the business students’ view with information
and communication technology students is interesting to study
because the later provide additional insight from information
technology literate consumers. Participation of business students
in this study is unlikely to present all the business students’ view;
therefore, similar study should be extended to cover as many as
possible business students in Malaysia higher education institu-
tions. Further, the negative and insignificant effect of perceive
size of trustees characteristics on consumer trust might be influ-
enced by the inability of respondents to imagine how big is the
online sellers’ market share. Therefore, it is important to objec-
tively show the market share of specific online sellers, so that,
respondents could make sound judgement on how perceive size
of the market share effect consumer trust.
References and Notes1. R. Connolly and F. Bannister, Management Research News 31, 339 (2008).2. T. S. H. Teo and J. Liu, Consumer trust in e-commerce in the United States,
Singapore and China, Omega (2007), Vol. 35, pp. 22–38.3. K.-H. Xiao, Journal of Advances in Information Sciences and Service
online shoppers in Malaysia. Available at: www.ecommercemilo.com/2014/01/ecommerce-infographic-malaysia-understanding-online-shoppers.html (2014).
5. H. van der Heijden, European Journal of Information Systems 12, 41.6. T. Li Wang and Y. F. Tseng, International Journal of Digital Society 2, 433
(2011).7. M. S. Md. Ariff, S. M. Yeow, and N. Zakuan, Adv. Sci. Lett. 20, 268 (2013).8. L. Zheng, M. Favier, P. Huang, and F. Coat, Journal of Electronic Commerce
Research 13, 255 (2012).9. S. C. Chen and G. S. Dhillon, Information Technology and Management 4, 303
(2003).10. Mastercard Worldwide Insights, Online shopping in Asia-Pacific-patterns,
trends and future growth, available at: www.mastercard.com/us/company/en/insights/studies/2008/asiaonlineshopping.htht (2008).
11. D. J. Kim and D. L. Ferrin, and H. R. Rao, Decision Support Systems 44, 544(2008).
12. D. G. Gambetta, Trust; Making and Breaking Corporative Relations, ElectronicEdition (1988), pp. 213–237.
13. S. Sahney, K. Ghosh, and A. Shrivastava, Journal of Asia Business Studies7, 278 (2013).
14. R. M. Morgan and S. D. Hunt, Journal of Marketing 58, 20 (1994).15. A. Mukherjee and P. Nath, European Journal of Marketing 41, 1173 (2007).16. Y.-H. Chen and S. Barnes, Industrial Management and Data Systems 107, 21
(2007).17. W. Gong, R. L. Stump, and L. M. Maddox, Journal of Asia Business Studies
7, 214 (2013).18. P. M. Doney and J. P. Cannon, Journal of Marketing 61, 35 (1997).19. C. Bianchi and L. Andrews, International Marketing Review 29, 253 (2012).20. I. Ajzen, Organizational Behavior and Human Decision Processes 50, 179
(1991).21. J. F. George, Internet Research 14, 198 (2004).22. P. Palvia, Information and Management 46, 213 (2009).23. Y. Xu and V. A. Paulins, Journal of Fashion Marketing and Management 9, 420
(2005).24. J. F. Hair, W. C. Black, B. J. Babin, R. E. Anderson, and R. L. Tatham,
Multivariate Data Analysis, 6th edn., Upper Saddle River, N. J., PearsonPrentice Hall (2006).
Received: 20 January 2015. Accepted: 20 February 2015.