基于支持向量机的多元文本分类研究  

Study on multiclass text categorization based on support vector machine

作　　者：魏新[1] 冯兴杰[1] 刘山[1] 

机构地区：[1]中国民航学院,天津300300

出　　处：《海军工程大学学报》2004年第5期30-32,42,共4页Journal of Naval University of Engineering

摘　　要：根据文本分类的特点,在对最小二乘支持向量机方法进行详细分析的基础上,创建了基于最小二乘支持向量机的多元文本分类器.实验表明,采用该文本分类器能够在保持较高分类精度和召回率的基础上,提高训练效率,具有一定的可行性.According to the particular property of text categorization, this paper carefully analyzes the theory of least squares support vector machine and constructs a multiclass text classifier based on least squares support vector machine. The experiments show that this classifier can improve the training efficiency and preserve the high precision and recall. This method of multiclass text categorization is workable.

关 键 词：支持向量机 文本分类 最小二乘支持向量机 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]