Presentation 2011-01-27
Text mining system STM based on semantic analysis
Minoru HARADA, Ryo ISHIDA, Kazuhiro YAMANISHI, Yoshiki KANDA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Current text mining based on the analysis on surface information of the morpheme and dependency among words cannot understand word meaning and functional relations between words, so it cannot find the knowledge concerning the relation that consists of two or more words such as "What is what" and "What does what do". Our text mining system STM analyzes sentences and phrases on the basis of the similarity of meaning of their content. In STM, Japanese sentences are converted into semantic graphs by our semantic analysis system Sage, and the degree of similarity of two sentences is measured based on the size of the similar common subgraph that consists of node pairs having similar meaning and arc pairs united by similar deep cases. As a result, the sentences with a similar meaning even if their expression are different are classified in a similar opinion cluster. In addition, the classification based on opinion person's feelings can be done by using the feelings word dictionary, the idiom dictionary, and the emoticon dictionary originally collected. Thus, we explain how a past text mining can be upgraded based on the semantic analysis.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Semantic analysis / Text mining / Feeling analysis Reputation analysis / Causality analysis / Sentence similarity
Paper # NLC2010-35
Date of Issue

Conference Information
Committee NLC
Conference Date 2011/1/20(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 Natural Language Understanding and Models of Communication (NLC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Text mining system STM based on semantic analysis
Sub Title (in English)
Keyword(1) Semantic analysis
Keyword(2) Text mining
Keyword(3) Feeling analysis Reputation analysis
Keyword(4) Causality analysis
Keyword(5) Sentence similarity
1st Author's Name Minoru HARADA
1st Author's Affiliation Faculty of Science and Engineering, Department of Integrated Information Technology, Aoyama Gakuin University()
2nd Author's Name Ryo ISHIDA
2nd Author's Affiliation Graduate School of Science and Engineering, Aoyama Gakuin University
3rd Author's Name Kazuhiro YAMANISHI
3rd Author's Affiliation Graduate School of Science and Engineering, Aoyama Gakuin University
4th Author's Name Yoshiki KANDA
4th Author's Affiliation Undergraduate School of Science and Engineering, Aoyama Gakuin University
Date 2011-01-27
Paper # NLC2010-35
Volume (vol) vol.110
Number (no) 400
Page pp.pp.-
#Pages 6
Date of Issue