基于支持向量机的网页多类分类技术

基于支持向量机的网页分类技术是数据挖掘中一个研究热点领域.支持向量机是一种高效的分类识别方法,在解决高维模式识别问题中表现出许多特有的优势,但支持向量机本身是一个两类问题的判别方法,不能直接应用于多类问题.总结了当前常用的几种支持向量机多类分类算法,分别从训练速度、测试速度、分类精度三方面对这些分类方法进行了讨论,并给出了进一步的研究方向.

Multi-sort classification of web page on SVM

Classification of web page based on the SVM was a hot point in data mining.SVM is an effective method learning the classification knowledge from massive data,especially in the situation of getting labeled classical examples costly.SVM can be used for binary-sort classification but not for multi-sort classification directly.Some methods for multi-sort classification based on SVMs were introduced and compared in three aspects: training speed,testing speed and classification accuracy.