Using linked open data to improve the search of open ... · Using linked open data to improve the search of open educational resources for engineering students Nelson Piedra, Janneth

Using linked open data to improve the search of open educational resources for engineering students

Nelson Piedra, Janneth Chicaiza, Jorge López Departamento de Ciencias de la Computación Universidad Técnica Particular de Loja, UTPL

Loja, Ecuador [email protected]; [email protected];

[email protected]

Edmundo Tovar Facultad de Informática

Universidad Politécnica de Madrid, UPM Madrid, España

[email protected]

Abstract— In this paper, authors apply the Linked Data

Design Issues to describe and retrieve information that is semantically related to open educational resources related to the Engineering Education, that are accessible via the OCW Higher Institutions. Linked data have the potential of create bridges between OCW data silos. To assess the impact of Linked Data in OCW, the authors present an interface of faceted search for open educational content. The authors demonstrate that OCW resource metadata related to engineering open courses can be consumed and enriched using datasets hosted by the LinkedOpenData cloud.

Keywords— OpenCourseWare; Open Educational Resources; OCW; OER; Linked Data; Faceted Search Engine; Serendipity; LOCWD

I. INTRODUCTION The key idea the Open Educational Resources (OER)

movement is that open educational content should be maximally shared [1]. Heterogeneity leads to problems of interoperability and accessibility of open content among institutions and within them. The lack of interoperability shows some disadvantages in the discovery, reuse, re-mix and adaptation of OER. OER Community must find a way to exchange quick and easy access to open educational materials.

Materials in OCW repositories are not usually described by metadata. Heterogeneity leads to problems of interoperability and accessibility of open content among institutions and within them. The lack of interoperability shows some disadvantages in the discovery, reuse, re-mix and adaptation of OCWs. OCW Community must find a way to exchange quick and easy access to OER.

Considerable work has been devoted to increase the interoperability between Learning Object Repositories that rely on different metadata schemas e.g. IEEE LOM. However, learning object metadata is typically not linked across repositories and not is possible navigate or interoperate between different data sources available on the Web. In this work, this problem is addressed through Linked Data by that describes how linked data has been integrated to data extracted from OCW repositories to navigate OCW resources.

Based on the perspective of Linked Open Data, free open OCW data also fosters interoperability and creates a basis on which the use, re-use, remix, and adaptation of open

educational tools or commercial applications can be built more easily. In Section 2, we describe the OpenCourseWare domain, and the notions of Linked Data and Linked OpenCourseWare Data Vocabulary, which ensures that an OER of type OCW can be safely discovered and reused. In Section 3, we present Serendipity, our implementation as well as an experimental evaluation of it. Finally, in Section 4, authors present the conclusions.

II. THE OPENCOURSEWARE DATASET

A. Data Source and Coverage There is not a standardized way to implement OCW

initiatives. The internal organization, structure and technological infrastructure of an OCW project are diverse, and respond to the vision of each university. Current OCW initiatives are Silos of OER. These silos of OER have no way to link to a particular item, and so hinder the free flow of information [8]. In this respect, OCW data is locked away in independent data silos, making it much less useful than it could be.

In this work, the data source is provided by higher education institutions associated to OCW Consortium (www.ocwconsortium.org/ ) and/or OCW – Universia (ocw.universia.net/).

Our dataset contains data about the main OCW concepts: (i) OER, the OER are teaching, learning, and research resources that reside in the public domain or have been released under an intellectual property license that permits their free use or re-purposing by others. (ii) OpenCourseWare, OCW is a type of OER.; (iii) repositories that contains OER and OCW courses; (iv) educational organizations; (v) users, creators or authors; (vi) branch of knowledge to which an OER belongs; and, (vii) open licenses or similar license that generally allows more liberal use, reuse, redistribution, re-mix and adaptation than a traditional copyrighted work.

B. Use of Linked Data from OCW content Semantic Web technologies and, more precisely, Linked

Data are changing the way information is stored and exploited. [5, 6]. The term “Linked Data” refers to a set of best practices for publishing and connecting structured data on the Web [4]. In summary, the Linked Data Design Issues, outlined by Tim Berners-Lee back in 2006, provide guidelines on how to use

978-1-4673-5261-1/13/$31.00 ©2013 IEEE

standardized Web technologies to set data-levdata from different sources [3, 7].

The motivation behind creating OpenCourseWare Data (LOCWD) is threefoldabout OCW resources such as those stored insystem are important for Educational dteachers, self-learners and other interested pparticularly interested in courses related to Table I and Table II). Secondly, although thereeducational datasets, none provide the leveLOCWD does, nor are they available as lithirdly, the current technology solutions useOER sources are limited to offering data vpages (not as raw data, but as embedded pages) and very few services through APIs.

The selected open educational content wLinked Data using the LOCWD vocabulary [9described in Linked Data/RDF were stored in this point, each resource was identified bydereferencing option, and thus display the reLinked Data.

TABLE I. OPEN COURSES RELATED TO ENGINE

University Massachusetts Institute of Technology, MIT

Universidad Carlos III de Madrid, UC3M Universitat Politécnica de Catalunya, UPC

The Open University, OU Universidad Politécnica de Madrid, UPM

Korea University, KU Universidad de Alicante, UA

Universidad Politécnica de Valencia, UPV

TABLE II. OCW - KNOWLEDGE AREAS RELATED

Knowledge Area related to Engineering Un

Electrical Engineering, Technological and Computer Sciences

MIT, UUPV, U

Mathematics & Physics MIT, UMechanical Engineering MIT, Chemical Engineering MIT, Materials Engineering UC3M

III. CURRENT USAGE It's difficult to develop tools for cons

multiple OCW silos. Searching OCW/OERsilos means invoking each one's user interfacthe results in separate groups. This severedevelopment of applications to OER/OCWcombine data from different sources. Howelinked data approach on OCW repositoriframework for its evolution into a more inintegrated system to sharing, connecting and and metadata of OCW initiatives.

The authors have personalized a faceted sconsumes data from LOCWD: Serendipity. ThSerendipity is based on flamenco.berkSerendipity we explore the potential of LOCWdiscovering process to give assistance to use,OER and OCW resources. See fig 1.

vel links between

the Linked d [9]: Firstly, data n the Serendipity

decision makers, persons. We are engineering (see

e are several open el of details that inked data. And, ed by OCW and via diverse Web data within web

were converted to 9]. The resources a RDF-Store. At

y a URI with a sults retrieved as

ERING (EXTRACT)

OCW quantity 704 148 95 87 56 38 38 33

D TO ENGINEERING

niversities UC3M, UPM, UPC, OU, KU,

UA UC3M, UPM UC3M, UPC UC3M, UPC

M, UPC, UPV

sume data from R across multiple

ce, and receiving ely impedes the

W that wish to ever, the use of es provides the nteroperable and discovering data

earch engine that he first version of keley.edu. With WD in supporting reuse and remix

A. Serendipity a Faceted Seach forSerendipity is an interface of fac

http://serendipity.utpl.edu.ec/. It isfrom OCW sites. Serendipity provallowing users to browse OpenCouway that they can rapidly get acqunature of the content, and never interface exposes OCW metadata inbuild their queries as they go, refininquery, with results automatically reThis interface also combines frcomplex search forms.

Faceted search, also called technique for accessing content faceted classification system, allocollection of information by applyfilters. A faceted classification information element along multenabling the classifications to be multiple ways rather than in taxonomic order.

Fig. 1. Explore OCW in an integrated anof the repositories of institutions that publish

B. Data and Facets in Serendipity In Serendipity, facets corresp

OpenCourseWare content.

Any of the following cases mstudent or self-learner to use SereUsers need to filter content ustaxonomy terms at the same time. text searches, category term filtering(iii) Self-learners don't know precisOCW site, or what to search for. clearly show users what subjecomprehensive on your site. (v) discover relationships or trends betwhas too much content for it to bnavigational structures, but you sti(vii) Self-learners want to use a facesingle taxonomic order or a single or sufficient for OCW content. (v

r OpenCourseWare Content ceted search accessible from s based on data extracted vides a search interface for urseWare content in such a uainted with the scope and feel lost in the data. This n such a way that users can ng or expanding the current eflecting the current query. ree-text search, it avoids

faceted navigation, is a organized according to a

owing users to explore a ying dynamic and multiple

system classifies each tiple explicit dimensions,

accessed and ordered in a single, pre-determined,

nd incremental manner, from any h OpenCourseWare.

ond to properties of the

might prompt to a teacher, entipity faceted search: (i) ing multiple category or (ii) Users want to combine g, and other search criteria. sely what they can find on (iv) Self-learners want to

ect areas are the most Self-learners are trying to

ween OCW. (vi) OCW sites be displayed through fixed ill want it to be navigable. eted classification because a folksonomy is not suitable

viii) Users often get empty

result sets when searching your site. (ix"advanced" search forms are not fun to use.

Fig. 2. Link from current OCW in Serendipity To Other

C. Navigation of OCW engineering courses In the search of OCW courses related

Serendipity demonstrates the following key fesearch results by facet, displaying a total numfacet value, refining search results by facet valfacet menu based on refined search criteria, dsearch criteria in a Bread Crumbs (navigationto exclude the chosen facet from the search cimprove ease of discovery open academic resimprove ease of consumption and reuse of reduce redundancy in search of OCW, and Cowith LOD data. Querying DBPedia, aadditional information about universities sudifferent languages, label, comment, latitudeFrom Geonames were extracted data aboucontinent, country, capital, city or state. Toexternal links, in this work were made diqueries (see Fig 2). With Serendipity, we dOCW resource data can be enriched using dathe LinkedOpenData cloud.

We have verified that the data published consistent and corresponds to the informatvarious OCW sites and OCW Consortiums.

IV. CONCLUSIONS The key idea the OER and OCW movem

educational content should be maximally sheducational data initiative should focus on

x) In cases that

LinkedData Source

to engineering, eatures: Grouping mber of OCW per lue, update of the displaying of the n guides), ability criteria, ability to sources, ability to OCW. ability to nnect OCW Data

authors obtained uch as name in e, and longitude. ut locations like o find and create irectly SPARQL

demonstrated that atasets hosted by

in Serendipity is tion obtained in

ment is that open hared. Any open n providing data

access permissions so that: praccess/use/reuse/create derivate worinformation and knowledge about th

A lot of open educational resouusage of open licenses, but not enstructure. Simply putting educationaopen license is obviously not enougresources data well depends on relright way. Moreover, with the prolifit has become increasingly hard to dhappening in Open Educational Con

In summary, we have shownprovides better results than other sihave also shown that our implemenlinked data technologies, can be mointeroperability and integration of O

ACKNOWLEDG

This work has been partially fuGovernment through the project 1650) ‘Research and Develtechnologies’, Universidad Técnicscholarship provided by the “SecretaSuperior, Ciencia y Tecnología” oFinally, authors thank the OCW Coand technical personal of TAWSBC

REFERENC

[1] Hilton, J., Wiley, D., Stein, J., & JohnOpenness and ALMS Analysis: FraResources. Open Learning: The Journa25(1), 37-44.

[2] Tuomi, I. (2006) Open Educational ResDo They Matter? Report for the OECDand Innovation. 5 October 2006.

[3] Bizer, C. (2009). The Emerging WSystems, IEEE, 24(5), 87 –92.

[4] Bizer, C., Heath, T., & Berners-Lee, TSo Far. International Journal on SSystems, 5(3), 1–22.

[5] Hausenblas, M. (2009). ExploitingApplications. Internet Computing, IEEE

[6] Bechhofer, S., Buchan, I., De RoureBhagat, J., Couch, P., et al. (2012). Wscientists. Future Generation Compute

[7] Passant, A., Laublet, P., Breslin, J. G.Worth a Thousand Tags. InternationaInformation Systems, 5(3), 71–94.

[8] Tovar, E., Piedra, N., Chicaiza, J., Lo(2012). OER Development and PromotResearch Projecton the OpenCourseWComputer Science, 18(1), 123–141.

[9] Piedra, N., Tovar, E., Colomo-Palaci(2013). Consuming and producing Opencourseware. Emerald EarlyCite,publication in 2013.

rovide non-discriminatory rks/adapt/share to raw data, he educational resource.

urces ends up focussed on nough attention is paid to al resources online under an gh. Doing open educational leasing key datasets in the feration of OCW/OER sites discovery and track what is ntext.

n that our implementation milar systems. Besides, we

ntation, which benefits from ore appropriate to deal with

OCW repositories.

GMENT unded by Madrid Regional eMadrid (Ref:S2009/TIC-opment for e-learning

ca Particular de Loja and aría Nacional de Educación of Ecuador (SENESCYT). onsortium, OCW-Universia

C on UTPL.

ES nson, A. (2010). The Four R’s of ameworks for Open Educational al of Open and Distance Learning,

sources: What Are They and Why D Centre for Educational Research

Web of Linked Data. Intelligent

T. (2009). Linked Data - The Story Semantic Web and Information

Linked Data to Build Web E, 13(4), 68 –73.

e, D., Missier, P., Ainsworth, J., Why linked data is not enough for

r Systems, in press. ., & Decker, S. (2009). A URI is al Journal on Semantic Web and

opez, J., & Martinez-Bonastre, O. tion. Outcomes of an International

Ware Model. Journal of Unviersal

ios, R., López, J. & Chicaiza, J. linked open data: the case of

, (PROG-Jul-2012-0045.R2), for