Class-based semantic cohesion computation

Author

Liu, Chang ; Zhou, Qiang

Author_Institution

Dept. of Comput. Sci. & Technol., Tsinghua Univ., Beijing, China

Volume

2

fYear

2003

fDate

5-8 Oct. 2003

Firstpage

1673

Abstract

Semantic cohesion is one kind of semantic relation between two words. It could be represented by the relation between the respective semantic meanings of the two words, and is useful for WSD problem in a collocation. In this paper, an HMM model was proposed to automatically label semantic meaning pairs for collocations composed of a noun and a verb. In order to find a suitable system of semantic class to represent individual word, three semantic classification methods based on the information from HowNet system, are proposed and compared with each other. The experimental results showed that one of the three classifications is appropriate for representing individual word. And these results also showed that the HMM model is helpful to captures the semantic cohesions properly.

Keywords

computational linguistics; grammars; hidden Markov models; semantic networks; unsupervised learning; HMM model; HowNet system; WSD problem; class based semantic cohesion computation; hidden Markov model; individual word; noun collocation; semantic classification methods; semantic meanings; semantic relation; unsupervised learning; verb collocation; word sense disambiguation; Clustering algorithms; Computer science; Hidden Markov models; Information processing; Information retrieval; Intelligent systems; Laboratories; Speech; Statistical analysis; Taxonomy;

fLanguage

English

Publisher

ieee

Conference_Titel

Systems, Man and Cybernetics, 2003. IEEE International Conference on

ISSN

1062-922X

Print_ISBN

0-7803-7952-7

Type

conf

DOI

10.1109/ICSMC.2003.1244653

Filename

1244653