Support Vector Machine Based on Hybrid Kernel Function

Abstract

Support vector machine use the kernel function to realize from the original input space to a high dimension space nonlinear mapping, and kernel function is the core of support vector machine, it is also the part which is difficult to understand of support vector machine. Because each of ordinary kernel functions has advantages and drawbacks, in order to get another kernel function with strong learning ability and generalization performance, this paper studies two kernel function of support vector machine—global kernel function(linear kernel function) and local kernel function(RBF kernel function), and presents combination kernel function of support vector machine. Through the experiment results comparing, results show that its performance is better than that of other SVMs constructed by ordinary kernel function.