关键词: v-支持向量机 核函数 泛化能力

Abstract:

Support vector machine has been effectively applied in pattern classification. For a sample space which is linearly inseparable,we need to find the ke  rnel function in order to map the sample set, which is linearly inseparable, to another high-dimension linear space. The effectiveness of the application of v-SVM is directly influenced by its generalization capability. In this article,we discuss an estimation of the generalization capability of v-SVM,  qualitatively analyze the influence on the generalization capability of v-SVM by the kernel function, and quantitatively analyze such influence by the G aussian kernel function.

Key words:  （v-support vector machine, kernel function, generalization capacity）