A new feature selection method based on distributional information for Text Classification

Feature Selection (FS) is one of the most important issues in Text Classification (TC). A good feature selection can improve the efficiency and accuracy of a text classifier. Based on the analysis of the feature´s distributional information, this paper presents a feature selection method named DIFS. In DIFS a new estimation mechanism is proposed to measure the relevance between feature´s distribution characteristics and contribution to categorization. In addition, two kinds of algorithms are designed to implement DIFS. Experiments are carried out on a Chinese corpus and by comparison the proposed approach shows a better performance.