关键词: 支持向量机, 核函数, 感知器, 核参数选取

Abstract: This paper proposes a kind of SVM classification，which is based on perceptron.This model in the classifier of training，has introduced the perceptron study thought，uses the support vector machines nuclear function to do nuclear calculation，then judges the classification of performance，classifying correctly does not make any revision，on the contrary，transforms to the perceptron study question.Experiments show that the model can not only improve the performance of SVM classification，but also can reduce the SVM classification performance to the nuclear function and parameters choice dependence.

Key words: Support Vector Machine（SVM）, Kernel function, perceptron, kernel parameter select