In view of word sense disambiguation shortcomings of the previous methods, they generally do not consider on word distance for computing semantic correlation of the influence of context, as well as the context is limited for ambiguous word sense disambiguation, and the use of part ambiguous context words make word senses more ambiguous. Therefore, this paper proposes the use of dependency parse tree and rules for feature word selection, then the ambiguous word is mapped to the Wikipedia pages to expand the feature words of the ambiguous word. Feature words expansion of word sense and feature words will eliminate the limitation of context words, by calculating improved mutual information between feature words of ambiguous context words senses and feature word of ambiguous word senses, then finally obtain the most suitable items of ambiguous word with context words in this sentence. Experimental results show that the proposed method compares with the previous method improves the accuracy of Chinese word sense disambiguation of 11.4%, with good scalability and practicality.
[1]
Razvan C. Bunescu,et al.
Using Encyclopedic Knowledge for Named entity Disambiguation
,
2006,
EACL.
[2]
Hou-Feng Wang,et al.
Chinese Word Sense Disambiguation Based on Maximum Entropy Model with Feature Selection: Chinese Word Sense Disambiguation Based on Maximum Entropy Model with Feature Selection
,
2010
.
[3]
Silviu Cucerzan,et al.
Large-Scale Named Entity Disambiguation Based on Wikipedia Data
,
2007,
EMNLP.
[4]
Jin Peng,et al.
Ensembles of Classifiers for Chinese Word Sense Disambiguation
,
2008
.
[5]
Wanxiang Che,et al.
LTP: A Chinese Language Technology Platform
,
2010,
COLING.
[6]
Marianna Apidianaki,et al.
Latent Semantic Word Sense Induction and Disambiguation
,
2011,
ACL.
[7]
Robert L. Mercer,et al.
Word-Sense Disambiguation Using Statistical Methods
,
1991,
ACL.