A Market-Based Recommender System
OA Location
Author(s)
Type
Conference Paper
Abstract
We have designed, implemented, deployed and evaluated a large-scale agent-oriented information system that recommends relevant documents to users. Our recommender system is now being used across several European institutions. Its two key features are a modular design capable of accomodating multiple recommendation methods, and the use of a marketplace to select and rank the best recommendations for the user. As part of our evaluation, we have extensively simulated this marketplace in order to understand its dynamics and validate its suitability for a recommender system.
Editor(s)
Giorgini, Paolo
Lespérance, Yves
Wagner, Gerd
Yu, Eric
Date Issued
2002
Citation
2002, pp.50-67
Publisher
http://CEUR-WS.org/Vol-59/
Start Page
50
End Page
67
Identifier
http://eprints.soton.ac.uk/257600/
Source
The Fourth International Bi-Conference Workshop on Agent-Oriented Information Systems at AAMAS 2002 (AOIS’02)
Notes
Event Dates: July 2002
Publication Status
Unpublished