摘要： 提出的最大散度差分类器是在修正Fisher线性鉴别准则的基础上建立起来的,它与Rocchio和SVM分类器有着十分密切的联系。在国际标准语料库20Newsgroups上进行的仿真实验结果表明,最大散度差分类器具有良好的文本分类性能,其正确识别率明显高于NaiveBayes和Rocchio,与SVM相当。

关键词: 最大散度差分类器, Naive Bayes分类器, Rocchio分类器, SVM分类器, 文本分类

Abstract: Maximum scatter difference classifier proposed in this paper is based on the modified Fisher linear discriminant. It has closed relationship with Rocchio and SVM. Experimental results conducted on 20 Newsgroups dataset show that MSD can achieve good performance to classify textual documents. Its recognition rate is significantly higher than that of Naive Bayes and Rocchio, and as high as that of SVM.

中图分类号: 

• TP391.1