GreenICT_Ontology
Post on 01-Nov-2014
104 Views
Preview:
DESCRIPTION
Transcript
Semi-automatic Green ICT Ontology Construction from CSR Report
by
Banatus Soiraya
Agenda
o Introduction & Motivationo Identification of Domain Structureo Ontology of Green ICT Domaino Results & Discussiono Conclusion
Effects of global warming
•Ecological and social changes
•The rise in global temperatures
Introduction
Warmer waters and more hurricanes Loss of biodiversity and animal extinction
Source : Wikipedia
Source : http://www.environmentalgraffiti.com/sciencetech/5-deadliest-effects-of-global-warming/276?image=17
Tsunamis
Introduction
• In fact, ICT industry accounts for 2 percent of Global CO2 Emissions equivalent to aviation industry. source : Gartner 2007
• For mitigating the issues, Going Green has become increasingly topic for enterprise companies in this decade.
• One of tools for going green is Green ICT.
Green ICT Definition
• Green IT definition : refers to environmentally sustainable computing or IT. source : wiki
• Green IT definition: Environment aspects and IT product and features. Source: TCO Certified
Green IT Framework
• Consists of 4 pillars and 5 rows. Source: Connection
Research
Green IT Scheme
• Studied CSR reports for content analysis and illustrating the relation among the entities. Source: Yulia et. al
Corporate Social Responsibility(CSR)
“Companies perform business activities whereas they sustain the environment for future populations.”Source : wiki
•Corporate Regulation Business Model•Enterprise Companies have to do CSR reports besides annual reports.
Ontology
• Ontology represents an important knowledge of specific domains and relationship.
Mobile Ontology source: Yu Zheng et. al
Motivation
• Many general terms of Green ICT.
• Not much appropriate forms or models of Green ICT
• Discover Green ICT knowledge from CSR report
• Aim to formalize GreenICT Ontology
Identification of Domain Structure (I)
Three steps
1. Collecting 45 CSR reports (.pdf) from websites as ranked by annual report 2007-2011 http://en.wikipedia.org/wiki/List_of_the_largest_technology_companies
2. Transforming .pdf to text by using : A-PDF Text Extractor and sobolsoft.com for removing line space
CSR Report
Pre-Processing
K-Mean
Weighting Terms
Filtering Terms
OntolgyDisplay
OntoGen
Identification of Domain Structure (II)
3. OntoGen: is a tool for:– N-Gram length– N-Gram Frequency– Stop word removing– Stemming – k Mean Clustering– Ontology Visualization
3. OntoGen: is a tool for:– k Means Clustering
Source: Ke Chen, The university of Manchester
Identification of Domain Structure (III)
Ontology of Green ICT Domain (I)
• Assigned 4 suggestions in OntoGen (4 number of cluster k)
• Four Categories are general terms when compare with corporate-social-responsibility-metrics.
Green ICT ontology with 4 Categories
Ontology of Green ICT Domain (II)
• Adding 4 suggestions in each category for constructing sub-categories.
• Sub-categories have more relation with corporate-social-responsibility-metrics.
1
23
4
1st Level 2nd Level
Green ICT ontology with 4 Categories and sub- Categories
Ontology of Green ICT Domain (III)
Green ICT ontology with 8 Categories
• Assigned 8 suggestions in OntoGen (8 number of cluster k)
• 8 Categories are general terms when compare with corporate-social-responsibility-metrics.
Ontology of Green ICT Domain (IV)
1
2
3
4
5
6
7
8
1st Level 2nd Level
Green ICT ontology with 8 Categories and sub- Categories
• Adding 4 suggestions in each category for constructing sub-categories.
• Sub-categories have more relation with corporate-social-responsibility-metrics.
• Note: Display only sub-categories that are different with main categories.
Results and Discussion (I)
Considering corporate-social-responsibility-metrics by Fujitsu and necessary metrics.
1.Employee health, safety and well being
2.Environmental (Co2 emission)
3.Energy
4.Travel
5.Community
6.Waste and Recycling
Results and Discussion (II)
• G1 stands for Green ICT Framework
• G2 stands for Green IT scheme
• G3 stands for Green ICT ontology with 4 categories
• G3-sub stands for Green ICT ontology with 4 categories and sub-categories
• G4 stands for Green ICT ontology with 8 categories
• G4-sub stands for Green ICT ontology with 8 categories and sub-categories
• G1 covers 6 criteria • G2 covers 5 criteria • G4-sub covers 5 criteria
Compare table of each model
Conclusion
• Sub-Categories will improve the efficiency of Green ICT Ontology.
• Green ICT ontology : 8 categories with sub categories will cover 5 criteria. It’s equal to Green IT scheme.
• Clustering techniques cannot identify the meaning of Green ICT or Grouping the same meaning Green ICT.
Future works
Focused AreasSemantic Clustering (Green ICT terms) Improving the accuracy of Green ICT Ontology
Techniques ConsiderationText miningNatural Language Processing
References
1. A Green ICT Framework Understanding and Measuring Green ICT , A Green Paper by Connection Research , April 2010.
2. Corporate Social Responsibility Metrics by Fujitsu, 2009.
3. Green IT – towards sustainability and a reduced carbon footprint. TCO Certified – White Paper, pp. 1-7.
4. http://en.wikipedia.org/wiki/Corporate_social_responsibility.
5. http://en.wikipedia.org/wiki/Green_IT.
6. Blaz Fortuna, Marko Grobelnik and Dunja Mladenic, OntoGen: Semi-automatic Ontology Editor, pp. 309-318, Springer-Verlag Berlin Heidelberg 2007.
7. Ronen Feldman, James Sanger, The Text Mining Handbook: Advanced Approaches in Analyzing Unstructured Data, CAMBRIDGE University Press, 2007.
8. Waralak V. Siricharoen and Thitima Puttitanun, Creating Ontology Chart Using Economy Domain Ontology Journal of Digital Content Technology and its Applications, 2009.
9. Yulia Wati and Chulmo Koo, The Green IT Practices of Nokia, Samsung, Sony and Sony Ericsson: Content Analysis Approach Proceedings of the 43 rd Hawaii International Conference on System Sciences, 2010.
10. Yu Zheng; Wenxiang Dou, Gengfeng Wu and Xin Li, Automated Chinese Domain Ontology Construction from Text Documents, pp. 639-648 Springer-Verlag Berlin Heidelberg 2007.