Presentation 2001/5/4
Automatic Segmentation of Compound Word in Japanese using Contextual Information
Dongli Han, Koichi Kato, Teiji Furugori,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Analyzing compound words is one of the crucial problems in constructing practieal natural language processing systems. In this paper, we propose a method for segmenting compound word, which consists of a long sequence of Kanji characters, in text by using statistics on word co-occurrences. We conducted an experiment that used the co-occurrence information within the compound word and the context in whieh it appreared. Its result shows a success rate of over 90% in dividing the compound words into their unit words.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Compound word / Segmentation / Co-occurrence / Mutual information / Contextual information
Paper # NLC2001-5
Date of Issue

Conference Information
Committee NLC
Conference Date 2001/5/4(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) Automatic Segmentation of Compound Word in Japanese using Contextual Information
Sub Title (in English)
Keyword(1) Compound word
Keyword(2) Segmentation
Keyword(3) Co-occurrence
Keyword(4) Mutual information
Keyword(5) Contextual information
1st Author's Name Dongli Han
1st Author's Affiliation Department of Computer Science, The University of Electro-Communications()
2nd Author's Name Koichi Kato
2nd Author's Affiliation Department of Computer Science, The University of Electro-Communications
3rd Author's Name Teiji Furugori
3rd Author's Affiliation Department of Computer Science, The University of Electro-Communications
Date 2001/5/4
Paper # NLC2001-5
Volume (vol) vol.101
Number (no) 40
Page pp.pp.-
#Pages 6
Date of Issue