支撑向量机的多类分类方法

Support Vector Machines for Multi-Class Classification

基于结构风险最小化原则的支撑向量机(SVM)具有良好的学习推广性,但是由于常规的SVM是从二类分类问题中推导出来的,在多类分类问题中就必须进行改进。文中讨论了支撑向量机的多类分类改进方法,运用在手写体数字识别中,并取得较好的结果。

Support vector machine is a new general machine learning tool based on structural risk minimization principle that exhibits good generalization .but it was originally designed for binary classification ,it must be improved when used in multi-class classification . this page discusses the multi-class classification based on support vector machine ,and applies it in the written digits recognition.