Presentation 2006-05-18
A Collaborative Filtering Method based on Marginal Rating Distributions
Shuhei KUWATA, Naonori UEDA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We propose a collaborative filtering method based on a probabilistic approach. In the proposed method, first empirical marginal distributions of the ratings over users and/or items are computed. Then, for unrated data, these distributions are represented as a function of unknown ratings, and the unknown ratings are predicted by minimizing KL divergence between both distributions. Unlike the conventional methods, the proposed method can predict all ratings simultaneously. In experiments using movie rating data, we confirmed that the proposed method could obtain significantly smaller prediction errors than the conventional probabilistic approach.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) collaborative filtering / marginal distribution / KL divergence / probabilistic approach
Paper # AI2006-3
Date of Issue

Conference Information
Committee AI
Conference Date 2006/5/11(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Artificial Intelligence and Knowledge-Based Processing (AI)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A Collaborative Filtering Method based on Marginal Rating Distributions
Sub Title (in English)
Keyword(1) collaborative filtering
Keyword(2) marginal distribution
Keyword(3) KL divergence
Keyword(4) probabilistic approach
1st Author's Name Shuhei KUWATA
1st Author's Affiliation NTT Comunication Science Laboratories, NTT Corporation()
2nd Author's Name Naonori UEDA
2nd Author's Affiliation NTT Comunication Science Laboratories, NTT Corporation
Date 2006-05-18
Paper # AI2006-3
Volume (vol) vol.106
Number (no) 38
Page pp.pp.-
#Pages 6
Date of Issue