Presentation 2004/7/7
An Automatic Selection Method of Integration Function
Yu SUZUKI, Kenji HATANO, Masatoshi YOSHIKAWA, Shunsuke UEMURA, Kyoji KAWAGOE,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Currently, many information retrieval systems that combine multiple retrieval results have been proposed to improve the accuracy of these infomration retrieval systems. In these systems, selections of the combination functions, the functions for combining multiple relevance scores, are important to improve the accuracies. We seem that the most suitable combination function should change by the users' queries. In this paper, we propose an automatic selection method of combination functions. In our method, we automatically select using the distributions of relevance scores. When a retrieval system calculates a small number of high relevance scores, the system can identify relevant retrieval targets. Therefore, we suppose that the sureness of the combination functions should depends on the number of high relevance scores. Then, we propose a calculating method of the sureness of combination function using Shannon's information measure. From our experimental results, we find out that our proposed method can select better combination functions.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Information Retrieval / Combination Function / Shannon's Information Measure / Image Retrieval
Paper # DE2004-81
Date of Issue

Conference Information
Committee DE
Conference Date 2004/7/7(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 Data Engineering (DE)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) An Automatic Selection Method of Integration Function
Sub Title (in English)
Keyword(1) Information Retrieval
Keyword(2) Combination Function
Keyword(3) Shannon's Information Measure
Keyword(4) Image Retrieval
1st Author's Name Yu SUZUKI
1st Author's Affiliation College of Information Science and Engineering, Ritsumeikan University()
2nd Author's Name Kenji HATANO
2nd Author's Affiliation Graduate School of Information Science, Nara Institute of Science and Technology
3rd Author's Name Masatoshi YOSHIKAWA
3rd Author's Affiliation Information Technology Center, Nagoya University
4th Author's Name Shunsuke UEMURA
4th Author's Affiliation Graduate School of Information Science, Nara Institute of Science and Technology
5th Author's Name Kyoji KAWAGOE
5th Author's Affiliation College of Information Science and Engineering, Ritsumeikan University
Date 2004/7/7
Paper # DE2004-81
Volume (vol) vol.104
Number (no) 177
Page pp.pp.-
#Pages 6
Date of Issue