AN EFFICIENT FEATURE SELECTION METHOD USING NAMED ENTITY RECOGNITION FOR CHINESE TEXT CATEGORIZATION

摘要

Feature selection Is an important task for test categorization. Traditional feature selection methods are based on terms but they may lose some useful information in texts. In this paper, we present a feature selection method that considers not only general terms but also named entities. Corresponding to our feature selection method, we propose a term weighting scheme for named entities. The experiments show that our method is effective comparing with traditional methods.