This paper presents an ”one fit all” solution for any field's text Word Sense Disambiguation(WSD), with a Sense Rank AALest algorithm derived from the Adapted of Lesk's dictionary-based WSD algorithm. AALesk brings a score for different relationship during gloss comparing, which makes WSD not only based on statistical calculate by process in a semantic way. Rather than simply disambiguate one word's sense one time, our solution considers the whole sentence environment and uses a Sense Rank algorithm to speed up the whole procedure. Sense Rank weights different sense combination according to their importance score. All these contribute to the accuracy and effective of the solution. We evaluated our solution by using the English lexical sample data from the SENSEVAL-2 word sense disambiguation exercise and attains a good result. Additionally, the independence of system components also make our solution adaptive for different field's requirement and can be easily improved it's accuracy by changing its core algorithm AALesk's parameter setting.
[1]
Hwee Tou Ng,et al.
Integrating Multiple Knowledge Sources to Disambiguate Word Sense: An Exemplar-Based Approach
,
1996,
ACL.
[2]
C. Fellbaum.
An Electronic Lexical Database
,
1998
.
[3]
Yaacov Choueka,et al.
Disambiguation by short contexts
,
1985,
Comput. Humanit..
[4]
George A. Miller,et al.
Using Corpus Statistics and WordNet Relations for Sense Identification
,
1998,
CL.
[5]
Christiane Fellbaum,et al.
Book Reviews: WordNet: An Electronic Lexical Database
,
1999,
CL.
[6]
Ted Pedersen,et al.
An Adapted Lesk Algorithm for Word Sense Disambiguation Using WordNet
,
2002,
CICLing.
[7]
Philip Resnik,et al.
Selectional Preference and Sense Disambiguation
,
1997
.
[8]
David Yarowsky,et al.
Unsupervised Word Sense Disambiguation Rivaling Supervised Methods
,
1995,
ACL.