基于SVM的多类分类算法改进

Improvement on bintree multi-class categorization algorithm based on SVM

在各种基于支持向量机的多类分类算法中,基于二叉树的多类支持向量机分类算法训练和分类速度相对较快,且解决了不可分问题,是一种很好的方法.本文系统研究和分析了基于二叉树的多类支持向量机分类算法,并在此基础上对其作出了改进,即当测试文本集规模较大时,对其先聚类再分类.改进的目的是,使测试文本不必总是从二叉树的根结点开始进行判断,而是有指导的代入分类函数中计算.在测试文本集规模较大,分类函数个数较多时,可以很大程度上增加分类效率,并加大了文本正确分类的概率.

It＇s a hotspot to research on support vector machine that extends from two-class issues to multi-class.Among all kinds of methods,bintree multi-class text categorization algorithm based on support vector machine is more effective in training and sorting than others,and it works out the impartibility problem.So it is a good method.The dissertation systematically researches and analyses bintree multi-class text categorization algorithm based on support vector machine,and then has some improvement on it.That is,we assembles firstly,and then sorts them when the size of testing texts is too large.The aim of this improvement is to make the testing text be computed more aimable,but does not begin from the base crunode of bintree at all time.The improvement can enhance the effect of text categorization and make it move accurat when the size of testing texts is too large and the quantity of sorted function is too much.