Please use this identifier to cite or link to this item:
Title: Scaling down candidate sets based on the temporal feature of items for improved hybrid recommendations
Authors: Tang, TY 
Winoto, P
Chan, KCC 
Issue Date: 2005
Publisher: Springer
Source: Lecture notes in computer science (including subseries Lecture notes in artificial intelligence and lecture notes in bioinformatics), 2005, v. 3169 LNAI, p. 169-186 How to cite?
Journal: Lecture notes in computer science (including subseries Lecture notes in artificial intelligence and lecture notes in bioinformatics) 
Abstract: The intensive information overload incurred by the growing interest in the Internet as a medium to conduct business has stimulated the adoption of recommender systems. However, scalability still remains an obstacle to applying recommender mechanism for large-scale web-based systems where thousands of items and transactions are readily available. To deal with this issue, data mining techniques have been applied to reduce the dimensions of candidate sets. In this chapter in the context of movie recommendations, we study a different kind of technique to scale down candidate sets by considering the temporal feature of items. In particular, we argue that movies' production year can be regarded as a "temporal context" to which the value (thus the rating) of the movie can be attached; and thus might significantly affect target users' future preferences. We call it the temporal effects of the items on the performance of the recommender systems. We perform some experiments on the MovieLens data sets. The results show that the temporal feature of items can not only be exploited to scale down the candidate sets, but also increase the accuracy of the recommender systems.
Description: IJCAI 2003 Workshop on Intelligent Techniques for Web Personalization, ITWP 2003, Acapulco, 11 August 2003
ISBN: 3540298460
ISSN: 0302-9743
EISSN: 1611-3349
Appears in Collections:Conference Paper

View full-text via PolyU eLinks SFX Query
Show full item record


Last Week
Last month
Citations as of Nov 13, 2018

Page view(s)

Citations as of Nov 11, 2018

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.