Download presentation
Presentation is loading. Please wait.
Published byNatalie Wood Modified over 9 years ago
1
Copyright 2009 Digital Enterprise Research Institute. All rights reserved. Digital Enterprise Research Institute www.deri.ie The SIOC MEthodology for Lightweight Ontology Development Uldis Bojārs uldis.bojars@gmail.com Tetherless World Constellation, RPI July 7, 2010
2
“In the Semantic Web, it is not the Semantics which is new, it is the Web which is new.” Chris Welty, IBM
3
Digital Enterprise Research Institute www.deri.ie The Web aspect of the Semantic Web Linked Data Social Semantic Web –Semantic Web + Social Web Lightweight Ontologies –Small vocabularies covering well-defined topics –Used in Semantic Web applications and mashups –Most ontologies on the Web are lightweight [d’Aquin et al., 2007] A Methodology for the Development of Lightweight Ontologies 3
4
Digital Enterprise Research Institute www.deri.ie Context The Web aspect of the Semantic Web Lightweight Ontologies Schema-like (Unpopulated) Ontologies The SIOC MEthodology –Ontology Design –Ontology Dissemination Case study: the SIOC Project 4
5
Digital Enterprise Research Institute www.deri.ie Lightweight / Heavyweight Lightweight ontologies include: –Concepts + taxonomies of concepts –Relationships between concepts –Properties that describe concepts Heavyweight ontologies: –Add axioms and constraints To clarify the intended meaning of the terms [Gómez-Pérez et al., 2003] [Mika and Akkermans, 2004] 5
6
Digital Enterprise Research Institute www.deri.ie Survey of Existing Methodologies General (application-independent): –Uschold and Gruninger Ontology capture, ontology coding, integration with existing ontologies –METHONTOLOGY Based on software engineering practices, uses evolving-prototype lifecycle Application-oriented: –On-To-Knowledge (feasibility study, baseline taxonomy, refinement) –eXtreme Ontology Methodology (test-driven) Do not consider the Web aspect of the Semantic Web 6
7
Digital Enterprise Research Institute www.deri.ie Survey of Existing Methodologies Heavyweight, expensive to use –steep learning curve; many modeling artifacts [Knoblauch, 2002] eXtreme Ontology Methodology is lightweight –but application-oriented (not applicable in a generic, app-independent case) –ignores the need to maintain and evolve the ontology eXtreme Ontology Methodology: –simplify existing ontology development methodologies –do a minimal set of changes the meet the requirements Uschold and Gruninger: –start by defining the purpose and scope On-To-Knowledge: –ontology maintenance activities –refining the ontology based on feedback 7
8
Digital Enterprise Research Institute www.deri.ie Problem 1: Ontology Design Existing methodologies –Focused on the development of heavyweight ontologies –Oriented towards KR experts –Expensive to use (steep learning curve) Usually not used by developers of lightweight ontologies Lightweight ontology development –Use different tools than heavyweight ontology developers –Processes are lighter and more informal –Development practices not studied in detail Target audience: –developers who are not necessarily ontology engineering or knowledge representation experts 8
9
Digital Enterprise Research Institute www.deri.ie Problem 2: Ontology Dissemination What makes [some] ontologies popular and widely used on the Web? –How can we influence & direct this outcome? Success stories: –FOAF, RSS 1.0 Useful, but not in wide use: –Annotea, TDL (Thread Description Language) Existing methodologies do not consider getting ontologies adopted on the Web 9
10
Digital Enterprise Research Institute www.deri.ie Proposed Solution “Streamlined”, lightweight development methodology –Including relevant guidelines: Development of Lightweight Ontologies Publishing on the Web Establish ontology promotion / dissemination as a core part of the ontology development process 10
11
The SIOC MEthodology
12
Digital Enterprise Research Institute www.deri.ie High-level view The SIOC MEthodology for Lightweight Ontology Development –Based on: Existing Methodologies (streamlined); Experience developing the SIOC ontology; Observations of development of lightweight ontologies on the Web Two main components: 12
13
Digital Enterprise Research Institute www.deri.ie Case Study: Semantically-Interlinked Online Communities (SIOC) 13
14
Digital Enterprise Research Institute www.deri.ie Ontology Design Process 14
15
Digital Enterprise Research Institute www.deri.ie Purpose & Scope, Choice of Guidelines What is the ontology built for? - How will it be used? –SIOC: communication, data integration Scope: defines the domain of the ontology –SIOC: blogs, forums, mailing lists (initial scope) Choice of Guidelines: –Ontology Language SIOC: OWL (initially: RDFS) –Versioning Approaches SIOC: ontology releases + revision control –Development Tools SIOC: text editor, SpecGen4, Pellet, Subversion, … Requirements –Core – Optional – Out-of-Scope 15
16
Digital Enterprise Research Institute www.deri.ie Requirements Analysis Analyse the domain area within the scope Describe the domain (“stories”) + a list of main terms Prioritise: core – optional – out of scope SIOC (blog information example): –“Blogs are a type of social media site” –“Blog posts usually have a title, creation date, …” –“Blog posts may be described using categories and tags” –Out of scope: post “mood” information smileys and icons that are sometimes associated with posts 16
17
Digital Enterprise Research Institute www.deri.ie Requirements (forums) 17
18
Digital Enterprise Research Institute www.deri.ie Modelling Keep It Small and Simple (KISS) – simple solutions Define Classes, Properties, Instances, Constraints Guidelines: –modelling strategy (top-down / middle-out / bottom-up) –naming conventions; documentation; modularity Ontology 101 [Noy and McGuinness, 2001] Case study: SIOC –Modelling strategy: middle-out –Naming conventions: UserAccount, has_creator, creator_of –Documentation: rdfs:label, rdfs:comment, per-term files –Modularity (4 modules – Access, Argument, Services, Types) 18
19
Digital Enterprise Research Institute www.deri.ie The SIOC Core Ontology 19
20
Digital Enterprise Research Institute www.deri.ie Ontology Reuse Identify Ontologies for Reuse –Sources: Semantic Web community; ontology search engines and libraries; publications and surveys; … Criteria for Reuse –Scope; Popularity; Documentation; Stability; … SIOC: –RSS 1.0 –Dublin Core –FOAF –SKOS 20
21
Digital Enterprise Research Institute www.deri.ie Ontology Reuse 21
22
Digital Enterprise Research Institute www.deri.ie Implementation Implement in the ontology language –RDFS, OWL Guidelines: choice of ontology development tools –Text editors; Web-based tools; IDE-type tools Case study: SIOC –Tools – text editor –Ontology language – OWL (a subset) –Serialisation – RDF/XML 22
23
Digital Enterprise Research Institute www.deri.ie Testing / Evaluation Test that the ontology, as implemented, is valid and consistent Check that the ontology matches its requirements Case study: SIOC –Test using RDF validator (RDF/XML), Pellet (OWL DL) –Check that the SIOC ontology satisfies its requirements 23
24
Digital Enterprise Research Institute www.deri.ie Design Evaluation: Req. for Blogs 24
25
Digital Enterprise Research Institute www.deri.ie Documenting, Publishing Specification of the ontology can be: –A) Generated from the ontology SIOC: Generated using SpecGen4 + templates + per-term docs –B) Use RDFa (same page = the ontology and its documentation) Other types of documentation Best practices for publishing ontologies –http://www.w3.org/TR/swbp-vocab-pub/ –Testing tools: Vapour –SIOC: http://rdfs.org/sioc/ns - using “#” namespacehttp://rdfs.org/sioc/ns separate documents for the ontology and its specification 25
26
Digital Enterprise Research Institute www.deri.ie Post-Development Activities –Dissemination of a New Release Maintenance and Evolution –Versioning; Modularity –Involving the Community Case study: SIOC –Versioning – ontology releases (current: 1.35) + Subversion –Evolution community-driven ontology terms were added, renamed, removed 26
27
Digital Enterprise Research Institute www.deri.ie Ontology Dissemination 27
28
Digital Enterprise Research Institute www.deri.ie Community Building and Involvement Attract a community around the project / the ontology –SIOC: http://sioc-project.orghttp://sioc-project.org Facilitate communication: mailing list, wiki, IRC, … –SIOC-Dev mailing list – 327 members, 1265 messages –#sioc IRC channel; SIOC wiki –SIOC meet-ups (brainstormings, …) Provide infrastructure –SIOC: Development server, Subversion, … (provided by DERI) 28
29
Digital Enterprise Research Institute www.deri.ie Marketing Activities Write about the ontology (information dissemination) –Follow related conversations on the Social Web Talks (academic, developer audiences) –Semantic Web conferences –“Unconferences”: BarCamp, WebCamp, VoCamp Feedback loops –SIOC: Aggregation on SIOC website, IRC bots Make information freely available –Presentations, Software –SIOC: SlideShare, … 29
30
Digital Enterprise Research Institute www.deri.ie SIOC: Marketing Publications –“Towards Semantically-Interlinked Online Communities” – 91 citation The 2nd European Semantic Web Conference (ESWC 2005) –“Interlinking the Social Web with Semantics” – 21 citation IEEE Intelligent Systems, 23(3): 29-40, 2008 Presentations / Tutorials –“Weaving SIOC into the Web of Linked Data” – 21,306 views Paper at the Linked Data on the Web workshop at WWW 2008 –“The Future of Social Networks on the Internet …” – 13,659 views Tutorial at the Semantic Technology 2008 conference Events –The Social Data on the Web workshop series (at ISWC 2008, 2009) “Tales from the SIOC-o-Sphere” summaries (1–9) by John Breslin 30
31
Digital Enterprise Research Institute www.deri.ie Using the Ontology Building Application Eco-System –“Bootstrap” the use of the ontology Cover the main parts of the information life-cycle –Data Creation –Data Integration –Using the Data http://rdfs.org/sioc/applications/ 31
32
Digital Enterprise Research Institute www.deri.ie Ontology Application Eco-System 32
33
Digital Enterprise Research Institute www.deri.ie SIOC in Use Bootstrapping: the 1 st SIOC applications –SIOC export plugin for Drupal –Semantic Radar for Firefox (originally: SIOCDetect) 23’867 downloads; ~2000 daily users –SIOC browsers, [Social] SIOC Explorer Built by the SIOC community: –Mailing List Archives, SWAML, Mailing List Explorer (MLE) –SIOC APIs (Java, Perl, …) –OpenLink DataSpaces –BlogEngine.NET –IkeWiki –Drupal 7 33
34
Digital Enterprise Research Institute www.deri.ie 34
35
Digital Enterprise Research Institute www.deri.ie Standardisation Collaboration with Standards Organisations –Benefit from reviewers’ feedback –Increase visibility and awareness –Collaboration with a team of project partners SIOC: W3C Member Submission (June 2007) –SIOC Core Ontology Specification –SIOC Ontology: Applications and Implementation Status –SIOC Ontology: Related Ontologies and RDF Vocabularies –Snapshots of RDF Namespace Documents SIOC Core Ontology + Types + Services modules 35
36
Digital Enterprise Research Institute www.deri.ie Evaluation: Dissemination of SIOC SIOC data on the Web: –Billion Triple Challenge 2009 dataset (vmlion25.deri.ie): more than 31 million RDF statements using SIOC properties more than 1.1 million instances of SIOC classes SIOC data in applications – 54 applications –21 made in DERI; 33 by the larger SIOC community –http://wiki.sioc-project.org/w/ApplicationTimelinehttp://wiki.sioc-project.org/w/ApplicationTimeline SIOC data in ontologies –Baetle (bug tracking ontology); myExperiment (science collab.); –MOAT (meaning of a tag); OPO (online presence); –SCOT (semantic tag-clouds); SWAN-SIOC (scientific discourse) 36
37
Digital Enterprise Research Institute www.deri.ie SIOC data on the Web: Classes Total: 1’110’665 –* in the BTC 2009 dataset 37 sioc:User709’942 sioc:UserGroup181’953 sioc:Post140’703 sioc:Forum24’351 sioc:Site22’457
38
Digital Enterprise Research Institute www.deri.ie SIOC data on the Web: Properties Total: 31’502’395 –* in the BTC 2009 dataset 38 sioc:links_to19’157 k sioc:has_container9’165 k sioc:name686 k sioc:topic670 k sioc:has_creator229 k
39
Conclusion
40
Digital Enterprise Research Institute www.deri.ie Summary Identified a problem Proposed a solution –The SIOC MEthodology –Dissemination – a core part of the development process Application Eco-System Evaluated via a case study –The design and dissemination of the SIOC ontology The case study (SIOC Project) is an important effort on its own and has impact outside the scope of the thesis –Enabling Inter-Operability on the Social Web 40
41
Digital Enterprise Research Institute www.deri.ie Future Work Development of Lightweight Ontologies –Study a larger range of ontologies Ontology Dissemination SIOC –Data Portability (in Practice) –Large-Scale Studies of Online Community Sites –New Kinds of Social Software 41
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.