文本分类中基于改进特征选择方法的研究  

Text Categorization Based on Improved Feature Selection in Text Categorization

作　　者：胡改蝶[1] 樊孝仁[1] 崔艺馨[1] 

机构地区：[1]太原工业学院网络与信息中心,太原030008

出　　处：《计算机与数字工程》2016年第7期1290-1292,共3页Computer & Digital Engineering

摘　　要：文本分类不仅可以提高分类的效率,而且可使人们更快地找到想要获取的信息。在特征选择方法的基础上,分析了卡方统计法的缺点,对其提出了一种改进的方法,同时采用支持向量机分类的算法和词频-逆向文件频率权重函数对其进行了验证。通过实验得出此方法可以在很大程度上提高文本分类精确度,使分类的效果更好。Text categorization not only can improve the efficiency of categorization,but also can make people quickly find the information they want.On the basis of the feature selection method,this paper analyzes Chi-square（CHI）statistical method shortcomings,and proposes a Chi-square statistical method.At the same time,the Support Vector Machine（SVM）classification’s algorithm and Term Frequency-Inverse Document Frequency（TF-IDF）weight function are used on the validation.The experiment shows that this method can largely improve to the text categorization accuracy,the classification effect is greatly improved,make better classification.

关 键 词：效率 文本分类 特征选择 卡方统计法 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]