Word sense disambiguation based on rough set

作     者：陈清才 王晓龙 赵健 陈滨 王长风 

作者机构：Dept. of Comp. Sci. and Eng. Harbin Inst. of Technol. Harbin 150001 China 

出 版 物：《Journal of Harbin Institute of Technology(New Series)》 (哈尔滨工业大学学报（英文版）)

年 卷 期：2002年第9卷第2期

页      面：201-204页

核心收录：

学科分类：081203[工学-计算机应用技术] 08[工学] 0835[工学-软件工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：SponsoredbytheNationalNaturalScienceFoundationofChina(GrantNo .60 175 0 2 0 )amdScientificFundforHeilongjiangOutstandingYouth(GrantNo .F0 2 0 60 4) 

主　　题：word sense disambiguation rough set sense feature system 

摘      要：A sense feature system (SFS) is first automatically constructed from the text corpora to structurize the textural information. WSD rules are then extracted from SFS according to their certainty factors and are applied to disambiguate the senses of polysemous words. The entropy of a deterministic rough prediction is used to measure the decision quality of a rule set. Finally, the back off rule smoothing method is further designed to improve the performance of a WSD model. In the experiments, a mean rate of correction achieved during experiments for WSD in the case of rule smoothing is 0.92.