基于改进互信息的特征提取的文本分类系统

Text Classification System Based on Feature Selection of Improvement Mutual Information

文章提出并实现了一种改进互信息的特征提取和支持朴素贝叶斯的文本分类系统,改进的互信息算法加强了负值单词的互信息值,弥补了原来互信息预处理算法的不足,从而提高了分类精度.实验结果表明本算法和系统具有较高的分类准确率。

This paper implements and take affect a kind of text classification system based on feature selection of improvement mutual information and supporting Naive Bayes, the algorithm of improved mutual information intensify the mutual information of negative, offset the deficiency of improved mutual information arithmetic and improve clasification precision, Experimental results indicate that algorithm and system have better clasification precision.