Support Vector Machine Based On Artificial Error

Content:Support vector machine(SV M)was introduced by Vapnik on the foundation of the statistical learning theory.In theory,SV M performs the structural risk minimization(SRM) principle,which minimizes an upper bound on the generalization error,as opposed to empirical risk minimization(ERM)which minimizes the error on training data.All these inherent characteristics decide that the generalization of SV M is good even in high dimensional spaces under small training sample conditions.Regarding classification,problem,when the training data set is linear,the basic idea of SV M is to Seek the optimal separating hyperplane for two kinds of samples in the original space;when the training data set is nonlinear,the basic idea of SV M is to map the input data into a higher dimensional feature space and then search for the optimal separating hyperplane in this feature space.Therefore,SV M resolves problems even if the samples is high-dimensional.Meanwhile, SV M tries to solve a quadratic programming problem with a linear constraint to seek a global optimized solution.This paper researches on support vector machine(SV M)in theory and model in order to solve classification problem.The following parts are main works:◆Obtains from two kinds of training samples approximate lin- earity separable situation,this article analyses and researches the traditional support vector machine algorithms and the deformation algorithms.◆In this paper,we analyse and study the traditional support vector machine algorithm and the deformation algorithm,considering the situation of the larger artificial error involved in training set data. Based on training data set by the larger impact of artificial error as the starting point,this article obtains from this flaw,to study,when the training regulations data included the artificial error how to safeguard the accuracy of the algorithms,proposes the support vector machines based on artificial error(artificial error—support vector machine hereinafter) thinking.◆This paper introduces the basic theory of the support vector machine based on the artificial error,and establishes the theoretical model of AE—SV M.This model is the improvement and the promotion of C—SV M model.