文本分类中特征选择方法的分析和改进  

Analysis and Improvement of Feature Selection for Text Categorization

作　　者：许朝阳[1] 

机构地区：[1]莆田学院电子信息工程系,福建莆田351100

出　　处：《计算机与现代化》2010年第4期37-39,共3页Computer and Modernization

摘　　要：随着WWW的迅猛发展,文本分类成为处理和组织大量文档数据的关键技术。基于向量空间的文本分类方法中,信息增益是一种有效的特征选择方法。本文改进信息增益的特征选择方法:降低负类对分类的贡献;改进特征项在类中的均匀程度对分类的贡献。实验结果表明,经过改进的信息增益的特征选择方法在分类效果上有显著的提高。With the development of World Wide Web,the text classification becomes a key technology in organizing and processing large amount of document data.Information gain is an effective feature selection in text classification based on vector space model.This paper improves the information gain feature selection： Reduce the negative contribution to the classification;improve the evenness of the feature in the class contribution to the classification.The experiments results show that the classification results can be improved significantly by improving the information gain feature selection.

关 键 词：文本分类 信息增益 特征选择 熵 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]