sociotm – relevancies, collaboration, and socio-knowledge in topic maps
DESCRIPTION
Topic Maps (TM) standard has solved a lot of problems in the information overload. With a semantic layer on the top of the existing data pools, TMs provide user-specific information interpretation and organization. However, one important component is missing in the TM standard and usage. This paper introduces SocioTM model; an extension of TM paradigm that includes relevancies, collaboration, and socio-knowledge (user-specific knowledge/behaviors) within TMs. The paper goes through relevancies implementation in SocioTM, relevancies building and population in SocioTM, relevancies interpretation, presentation, and navigation through SocioTM. Relevancies are presented on two levels: topic/ontology level and information (occurrences) level. The paper concludes with collaboration involvement in SocioTM building and with migration of socio-knowledge.TRANSCRIPT
Саша Рудан
Relevancies, Collaboration, and Socio-knowledgeRelevancies, Collaboration, and Socio-knowledgein Topic Mapsin Topic Mapsin Topic Mapsin Topic Maps
([email protected])Синаша Рудан
Socio TMSocio TM
Introduction
2/16www.SocioTM.com, TMRA 2008, Leipzig
www.KnowledgeFederation.com
www.SocioTM.com Poliscopy
Knowledgeintegration
www.MemePolis.comwww.KnAlledge.com www.Fuzzzy.com
www.HeadWareSolutions.comwww.MagicWandSolutions.com
Independentdevelopers
• Users are interested in– Browsing in meta-data
(knowledge) space• not information space
– Topics that are more relevant
• Users are looking for– not only knowledge but also– something cool, interesting,
exciting, unknown
• » la science pour la science «– Knowledge Isolation
• TM knowledge is hard-typed, it is not fuzzy
User/Our Objectives
3/16www.SocioTM.com, TMRA 2008, Leipzig
• Ubiquity phenomena• No talk about recommenders
– Google news, Amazon, MovieLens, etc
• Ranking topics of – the same class– different classes
Relevancies, Ranking, Recommenders
4/16www.SocioTM.com, TMRA 2008, Leipzig
• Ranking support in TM– Scopes– Associations
• Problems– Binary-like concepts– Hardcoded in ontology space
• More fuzzy and general concept is needed
State of The Art
5/16www.SocioTM.com, TMRA 2008, Leipzig
1. Information and meta-data space huge and highly interconnected
2. Interconnections important for user3. No need for real-time knowledge rebuilding4. Need for structural concept of knowledge5. Users:
a) Need data ranking and data recommendationsb) Need to individually affect knowledge andc) Preferably even globallyd) Able to migrate with aggregated socio-knowledge
Problem setting
www.SocioTM.com, TMRA 2008, Leipzig6/16
SocioTM modelSocioTM modelOur approachOur approach
• Relevancies on – Occurrence and
– Topic map constructs• Behavioral prediction• Socio-knowledge migration
• Global knowledge structure building
• Unique vs. global knowledge• Mountain view
• View-clipping7/16www.SocioTM.com, TMRA 2008, Leipzig
Global and Conceptual overview
SocioTM Model (continued)SocioTM Model (continued)
www.SocioTM.com, TMRA 2008, Leipzig8/12
• Implementation• User implicit feedback• Relevance evolution
Relevance population and creation
www.SocioTM.com, TMRA 2008, Leipzig9/16
• fUSER-NORMALIZING = fGROUP-PROFILE x fUSER-PROFILE
x fST/LT-INTERESTS x fSEARCH-ITEM
x fEXPLICIT-SOCIO-KNOLEDGE x fNAVIGATION x fTIME
Relevance Evolution/Normalizing
www.SocioTM.com, TMRA 2008, Leipzig10/16
• Building dynamic SocioTM– As copy of topic space– On-the-fly
• Wide normalization– User profile– Short-term user interests– Search-item– User explicit socio-knowledge
• Normalizing through navigation and time
SocioTM Interpretation
www.SocioTM.com, TMRA 2008, Leipzig11/16
• Challenges– Overloading– Navigation– Understanding
• View-clipping• Mountain-view paradigm
SocioTM Presentation
www.SocioTM.com, TMRA 2008, Leipzig12/16
Socio-Potential Law
www.SocioTM.com, TMRA 2008, Leipzig13/16
• Blue Circle – search location• Red Ring – browsing location• a) - initial position, b) after fist navigation step• a1, b1) - non-evaluated, a2, b2) – evaluated
• Socio-knowledge migration• Needs for pre-populated user preferences for
a new TM• Needs for global understanding user behaviors
Collaboration within Topic Maps
www.SocioTM.com, TMRA 2008, Leipzig14/16
• Introducing SocioTM into the existing systems
• Vertical compatibility• Standardizing interface• Proxy implementation
SocioTM Implementation
www.SocioTM.com, TMRA 2008, Leipzig15/16
• Extend/add to TM standard • not just a business-logic module• Introducing fuzzy, multy-view knowledge• Living, user-affected knowledge
Conclusion
www.SocioTM.com, TMRA 2008, Leipzig16/16
• There is no standard defect but it is extension• Should we bother other users with our
opinions?• Why it should be standard and not just
business logic on the top of it?
Opened (?) Questions
www.SocioTM.com, TMRA 2008, Leipzig17/16
Саша Рудан
Relevancies, Collaboration, and Socio-knowledgeRelevancies, Collaboration, and Socio-knowledgein Topic Mapsin Topic Maps
SocioTMSocioTM
([email protected])Синаша Рудан
T:hank you!