Survey of multi-classification algorithms based on support vector machine

The basic theory and algorithms of statistical learning theory(SLT) and support vector machines(SVMs) are surveyed.According to 2 classification and multi classification cases,main SVM training algorithms are summarized and compared.By comparison with ANN,the characteristics of SVMs are analyzed.SVM applications,such as pattern recognition,function approaching,time series prediction,fault prediction and recognition,information security,power system and power electronics,are described.Finally,some problems in SVM development are presented.