基于支持向量机的文本分类方法

Text categorization method based on SVM

自动文本分类是信息处理的主要研究内容.分析中文文本分类的基本过程和主要技术,并用支持向量机(SVM)实现文本分类.比较多项式、径向基和Sigmoid核函数的分类效果.结果表明,将中文自然语言平台的语料库中1900个文本作为测试样本和训练样本,采用径向基核函数的SVM分类算法,实现中文文本分类的效果最佳,总准确率达到88.579%.

Text automatic classification is the main part of information processing.This paper introduces the basic process and main techniques of Chinese text classification,and makes the text categorization system based on Support Vector Machines(SVM)come true.We compared the text classification results with three kinds of kernel functions of SVM.The experiments show that RBF kernel function can achieve the best result.