摘要： 基于SVM理论的分类器已经发展为一种通用的二值分类器,但它不适用于多值的场合。该文在分析经典的SVM分类算法和决策树分类算法的基础上,采用了将SVM和二叉决策树结合的方法来实现多类问题的分类,实验表明它与一对一的多值分类器具有相似的分类精度,并比一对一的分类算法具有快的测试速度。

关键词: 支持矢量机, 二叉决策树, 多值分类器

Abstract: Classifier based on SVM theory has been developed as a general-purpose two-class classifier,but it is not suitable for multi-classclassification.Two basic classification algorithms are analyzed and combined into one multi-class classifier.The experimenttal results show that it isgood as ‘one against one’ algorithm in classification preciseness and it has much higher test speed.

中图分类号: 

• TP18