Arghtoo much

1
Arghtoo much

• Bijan Parsia

2
Agenda

• FilmTrust
• Recommender systems
• Swoop
• Ontology browsing and editing and annotating and
  debugging and engineering
• How we make our research usable
• Econnection based ontology partitioning
• (No services, for right now)

3
FilmTrust

• http//trust.mindswap.org/FilmTrust
• An experiment in trust network based inference
  support for recommender systems
• Part of Jen Golbecks doctoral work
• Join!
• Go to FilmTrust
• Join
• Rate some films
• Link to someone else!!!
• I.e., have friends get them to join

4
FilmTrust Recommendations

• Weight ratings by trust value
• Rating r from node s to movie m
• E.G.
• Alice rates Jaws at 3 stars
• Bob rates Jaws at 1 star
• Chuck trusts Alice at level 8
• Chuck trusts Bob at level 2
• The recommended rating for Chuck
• 83 21 / (82) 26/10 2.6 Stars

5
Accuracy of Recommendations

• Overall, users make ratings that are close to the
  average, so overall the average is a good
  predictor
• When the users rating differs from the average,
  the trust-recommendation
• Outperforms the average
• Outperforms the standard correlation-based
  collaborative filtering algorithm

6
(No Transcript)
7
Swoop

• Web Browser inspired ontology browser
• Lots of editing support
• Debugging
• Partitioning
• Search
• Making our research usable and accessible
• (For us too!!!)
• OnToDemo