Home // SEMAPRO 2014, The Eighth International Conference on Advances in Semantic Processing // View article

Word Sense Disambiguation Based on Semi-automatically Constructed Collocation Dictionary

Authors:
Minoru Sasaki
Kanako Komiya
Hiroyuki Shinnou

Keywords: word sense disambiguation; one sense per collocation; sense-tagged collocation dictionary construction

Abstract:
In this paper, we propose a novel Word Sense Disambiguation (WSD) method based on collocation that has a particular meaning. This proposed method is to identify the sense of idiom or common phrase containing a target word before the existing statistical WSD method is applied by capturing the context information. To evaluate the efficiency of the proposed method using collocation dictionary, we make some experiments to compare with the result of the Support Vector Machine (SVM) classification. The results of the experiments show that almost the sense of the extracted collocation has only one particular sense when we obtain the word pair of (the target word, noun word) and (noun word, the target word) with high pointwise mutual information value. Moreover, in the experiment of WSD task, the total average precision of our system is improved in comparison with the baseline system using SVM.

Pages: 29 to 32

Copyright: Copyright (c) IARIA, 2014

Publication date: August 24, 2014

Published in: conference

ISSN: 2308-4510

ISBN: 978-1-61208-355-1

Location: Rome, Italy

Dates: from August 24, 2014 to August 28, 2014