Support Vector Machine (SVM) In The Relevant Parameter Selection And Application

Support vector machine(SVM)that based on the framework of the statistical learning theory is a new kind of algorithm , and it has obvious advantages in dealing with classification problem.At present,the research on the selection of the related parameter of support vector is increasing ,and it become more maturely,but the unified method of the selection of related parameter hasn't been formed,in most cases,only rely on experience or contrast can obtain.The purpose of this paper is to study the selection and the applying of Gaussian kernel parameter ? and penalty parameter ? in support vector machine.In the view of this situation,the work is as follows:Firstly ,the paper introduces the background, significance and current situation of the research,and also introduces the basic knowledge of support vector machine (SVM) to pave the way for the following.Secondly, the basic principle of particle swarm optimization (pso)algorithm is introduced.Through the correlation analysis,an improved particle swarm optimization (pso) algorithm has been proposed in theory,and this algorithm is applied to select the parameters of Gauss kernel and penalty parameters. Through the comparison with the standard particle swarm algorithm and fundamental particle swarm algorithm in numerical experiment, proves that the presented algorithm has better classification effect;Finally,the nuclear parameter that been calculated in improved particle swarm optimization will be applied to the face recognition classification.Applied the new algorithm to the classifier that constructs Kernel function of support vector machine (SVM),and compare it with the classifier about the standard particle swarm algorithm and fundamental particle swarm algorithm on the training time and recognition rate,shows that the algorithm has superiority in classification.