Virtual Community Modelling to Support Knowledge Sharing

Document Sample
Virtual Community Modelling to Support Knowledge Sharing Powered By Docstoc
					          Computing
School of something
           ENGINEERING
FACULTY OF OTHER




    Virtual Community Modelling to Support
              Knowledge Sharing




Kleanthous Styliani
www.comp.leeds.ac.uk/stellak
                               Supervised by: Dimitrova Vania
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work

  Research Focus:
  Provide holistic personalised support in VC


  What is a Virtual Community?
  People using computers and the internet to communicate and share interests
  & knowledge

     Loosely Structured                           Closely-Knit
                                                    •Educational
                                                    •Organisational



Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work


  Chracteristics of Closely Knit VC…


     • Controlled Membership   • Sharing information
     • Common Purpose          • Generation of new
                                 Knowledge
     • Shared Interests
                               • Collaboration




Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work




Duplication of resources
                             Ignorance of others knowledge

                        +
                                    New members
                                Fischer & Ostwald (2001)not   integrate


 Active members becoming inactive

Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work

  Our Approach…
  Develop intelligent techniques to automatically detect required
  support tailored to the community
      User Modelling Approaches




            VC as an entity


  What processes are important for the effective functioning
     of a community and how can they be supported?
Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work

•Shared Mental Models
•Transactive Memory
•Cognitive Centrality




Convert Data
                  Ontology                           Paper at UM 2007 Conference

   Kleanthous, S. and Dimitrova, V. (2007) A Semantic-Enhanced Approach for Modelling
   Relationships and Centrality in Virtual Communities, SociUM: 1st Workshop on Adaptation
   and Personalisation in Social Systems: Groups, Teams, Communities at 11th International
   Conference on User Modeling, Corfu, Greece

Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work


 Community Model Application…
  Relationship model extracted as
  graphs

  ReadSim, UploadSim and
  InterestSim: Undirected graphs
  ReadRes: Directed graph

  Problems Identified


                                                       Paper at AH 2008 Conference
 Kleanthous, S. and Dimitrova, V. (2008) Modelling Semantic Relationships and Centrality to Facilitate
 Community Knowledge Sharing, 5th International Conference on Adaptive Hypermedia & Adaptive
 Web-Based Systems (AH'08), Hannover, Germany

Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work


  Automatic graph-based pattern detection…
  P1: Two members have a <relationship> with
  the same members but not among themselves


                                                    Importance
ReadSim for members a and b:
                                                    •Improve TM System
    GRS (VRS , ERS )    ΝRS (v a )   ΝRS (v b )
                                                    •Promote Collaboration


       ΝRS (v a )  ΝRS (v b )     v a  ΝRS (v b ) 

Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work



 P2: A member who appears uploading but
 not downloading and others are reading what
 he/she is uploading.                              Importance
 For member a :                                    •Motivate to start
                                                   downloading
 ReadRes          GRR (VRR , ERR )                 •Make member aware of
                                                   similar members
                    
                   ΝRR (v a )                      •Develop SMM
                                                   •Collaboration

     uRate(a)  0    dRate(a )  0   ΝRR (v a )   
                                             




Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work


 P3: A member appears to download only and
 has InterestSim with other members.

                                                              Importance
 For member a :
                                                              •Motivate uploading
 InterestSim GIS (VIS , EIS )                                 •Development of SMM
                    ΝIS (v a )                                •Improve TM system
                                                              •Collaboration



   uRate(a )  0    dRate(a )  0   ΝIS (v a )   


Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work

 Algorithm application & Results…
    Study from October 2005 – Dec 2006
  Results from study October 2005 – May 2006
                                 140

   Barriers for:                 120

   •Development of TM system     100

   •Building of SMM              80

   •Collaboration not possible   60

                                 40

                                 20

                                  0
 Results from study January,           Oct- Nov- Dec- Jan- Feb- Mar- Apr- May- Jun- Jul- Aug- Sep- Oct- Nov- Dec-
                                        05   05   05   06   06   06   06   06   06   06   06   06   06   06   06
 February and March 2006                                         Downloading     Uploading

 •Each month closer to stop functioning
 •More problems, identified on March 2006 than on January 2006

Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work



    January 2006: Members 9 and 24 had ReadSim relation with the
    same members as member 5 but not with member 5.


    February 2006: Members 9 and 24 disengage from the community



  Support
  •Automatic messages to members 9 and 24 to make them aware of their
  similarity to 5 and others.

  •development of a better TM system
  •open the doors for collaboration


Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work


  January 2006: member 28 appears to be a connecting node between
  members 5, 9, 24 and 31. All four members appear to have ReadSim with
  member 28 but not among themselves.



 Support
 •Make 28 aware of his importance to the community
 •Encourage 28 to collaborate with members 5, 9, 24 and 31
 •Make 5, 9, 24 and 31 aware of their connections through 28


 •Facilitate SMM
 •Improve TM System
 •Support Knowledge Sharing

Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work

   January 2006: one member was only uploading and four members were only
   downloading.
   March 2006: two members were only uploading and nine members only
   downloading.
   e.g. Member 19 downloaded thirty-three resources, without uploading
   anything from January to March

  Support
  • provide information of members with similar interests based on InterestSim
  • members who are reading similar resources (ReadSim)
  • uploading resources similar to member’s interests (UploadSim).

  •Motivate members to contribute
  •Improve SMM, TM
  •Facilitate knowledge sharing
Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work


   Future Work…
   Algorithms for detecting community changes
     implemented & applied to VC
     • Analysis of results (currently working)

   Final Evaluation Study applying the whole framework
     • Derive a community model
     • Detect problematic situations
     • Intervene to improve community functioning
     • Analyse community changes after adaptive intervention
     • Analyse subjective data

Nov 5th 2008
  Virtual Communities
  PhD Overview
  KSBP Algorithms
  Results
  Future Work

  Summary…
  •Problem: There is a need to support VC
  •Solution: Intelligent techniques tailored to the whole community can
  provide the foundations for a sustainable VC
  •Results: TM, SMM, CCen can be used to support VC
        • Modelling semantic-enhanced relationships can help us to identify
          what support is needed
        • Automatic pattern detection can help in generating adaptive
          interventions
  •Future Work
        • How changes in the VC over time can influence knowledge sharing?
        • Evaluation of the approach on a different VC

Nov 5th 2008