Capturing Interest Through Inference and Visualization: Ontological User Profiling in Recommender Systems
Capturing Interest Through Inference and Visualization: Ontological User Profiling in Recommender Systems
Tools for filtering the World Wide Web exist, but they are hampered by the difficulty of capturing user preferences in such a diverse and dynamic environment. Recommender systems help where explicit search queries are not available or are difficult to formulate, learning the type of thing users like over a period of time. We explore an ontological approach to user profiling in the context of a recommender system. Building on previous work involving ontological profile inference and the use of external ontologies to overcome the cold-start problem, we explore the idea of profile visualization to capture further knowledge about user interests. Our system, called Foxtrot, examines the problem of recommending on-line research papers to academic researchers. Both our ontological approach to user profiling and our visualization of user profiles are novel ideas to recommender systems. A year long experiment is conducted with over 200 staff and students at the University of Southampton. The effectiveness of visualizing profiles and eliciting profile feedback is measured, as is the overall effectiveness of the recommender system.
Knowledge capture, Machine learning, Ontology, Profile visualization, Recommender systems, User profiling, User modelling
Middleton, Stuart
404b62ba-d77e-476b-9775-32645b04473f
Shadbolt, N.R.
5c5acdf4-ad42-49b6-81fe-e9db58c2caf7
De Roure, D.C.
02879140-3508-4db9-a7f4-d114421375da
2003
Middleton, Stuart
404b62ba-d77e-476b-9775-32645b04473f
Shadbolt, N.R.
5c5acdf4-ad42-49b6-81fe-e9db58c2caf7
De Roure, D.C.
02879140-3508-4db9-a7f4-d114421375da
Middleton, Stuart, Shadbolt, N.R. and De Roure, D.C.
(2003)
Capturing Interest Through Inference and Visualization: Ontological User Profiling in Recommender Systems.
K-CAP2003, Sundial Beach Resort, Sanibel Island, Florida, United States.
Record type:
Conference or Workshop Item
(Paper)
Abstract
Tools for filtering the World Wide Web exist, but they are hampered by the difficulty of capturing user preferences in such a diverse and dynamic environment. Recommender systems help where explicit search queries are not available or are difficult to formulate, learning the type of thing users like over a period of time. We explore an ontological approach to user profiling in the context of a recommender system. Building on previous work involving ontological profile inference and the use of external ontologies to overcome the cold-start problem, we explore the idea of profile visualization to capture further knowledge about user interests. Our system, called Foxtrot, examines the problem of recommending on-line research papers to academic researchers. Both our ontological approach to user profiling and our visualization of user profiles are novel ideas to recommender systems. A year long experiment is conducted with over 200 staff and students at the University of Southampton. The effectiveness of visualizing profiles and eliciting profile feedback is measured, as is the overall effectiveness of the recommender system.
Text
kcap2003.pdf
- Other
More information
Published date: 2003
Additional Information:
Event Dates: October
Venue - Dates:
K-CAP2003, Sundial Beach Resort, Sanibel Island, Florida, United States, 2003-10-01
Keywords:
Knowledge capture, Machine learning, Ontology, Profile visualization, Recommender systems, User profiling, User modelling
Organisations:
Web & Internet Science, Electronics & Computer Science, IT Innovation
Identifiers
Local EPrints ID: 258932
URI: http://eprints.soton.ac.uk/id/eprint/258932
PURE UUID: 48ef7da2-bdd3-4d1a-8a37-0f8814a8dbbd
Catalogue record
Date deposited: 05 Mar 2004
Last modified: 15 Mar 2024 03:08
Export record
Contributors
Author:
N.R. Shadbolt
Author:
D.C. De Roure
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics