基于两种特征贡献度的特征选择  

Feature Selection Based on Two Kinds of Feature Contribution Degree

作　　者：周瑞琼[1] 朱颢东[2,3] 吴洪丽[1,3] 

机构地区：[1]海南师范大学信息科学技术学院,海南海口571158 [2]郑州轻工业学院计算机与通信工程学院,河南郑州450002 [3]中国科学院成都计算机应用研究所,四川成都610041

出　　处：《河南科技大学学报（自然科学版）》2010年第5期44-47,共4页Journal of Henan University of Science And Technology:Natural Science

基　　金：海南省自然科学基金项目(807050);四川省科技计划项目(2008GZ0003)

摘　　要：特征选择是提高文本分类性能的一个重要手段。论文首先定义了两种特征贡献度:一个是特征对类间文档分散程度的贡献度,该贡献度越大越好;另一种是特征对类内文档分散程度的贡献度,该贡献度越小越好。然后把这两种特征贡献度有机地结合起来设计了一个新的特征选择方法,该方法能够对所选特征进行综合考虑,从而使获得的特征集具有较好的代表性。仿真实验表明:所提特征选择方法在一定程度上能够提高文本分类性能。Feature selection is one of important means that improves performance of text categorization.Firstly,two kinds of feature contribution degree were defined.One was the feature contribution degree of document dispersion degree between categories,the larger the better.The other was the feature contribution degree of document dispersion degree in category,the smaller the better.And then,the two kinds of feature contribution degree were integrated and a new feature selection method was designed.The method can inspect selected feature synthetically so that the more representative feature set is obtained.Simulation experiments show that,to a certain extent,the proposed feature selection method is able to improve performance of text categorization.

关 键 词：特征选择 文本分类 特征贡献度 文档分散程度 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]