一种改进的文本特征选择方法

随着Internet技术的飞速发展，网页上存在着各种各样、类目繁多的信息，因此网页分类技术就显得越来越有意义。本文使用向量空间模型（VSM）来表示网页文本，提出了一种改进X2的文本特征选择方法，最后通过支持向量机方法进行分类。实验结果表明，相对于传统的X2文本分类统计方法，改进后的特征选择方法的分类效果要好于传统的X2统计方法。

An improved text feature selection method

With the rapid development of Internet technology,there are many categories of information on the Web.Therefore the technology of Web page classification becomes more and more meaningful.Using the vector space model(VSM) to indicate the page text.A feature selection method is presented to improve χ2statistical methods.Finally support vector machine is usd to classify.Through a text classification experiment,the results show that compared with traditional χ2 statistical methods,the improved feature selection method is better than traditional χ2 statistical methods.