Comparison of Multi-class SVMs Methods

Statistical Learning Theory(SLT),the first theory that systematically studies the problem of machine learning with small size samples,presents a new inductive principle,structural risk minimization(SRM) principle,which tells us how to select the suitable classification model according to sample amount so as to obtain high generalization ability.Support vector machine(SVM) is a new commonly used machine learning method based on SRM.Some suggestions to the improvement of the method were provided.The systematically analysis of the existing multi-class SVMs(M-SVMs) methods shows that hierarchy multi-class SVMs,BT-SVMs can be relatively effective in multi-class classification.