並列摘要
Support Vector Machine is a kind of algorithm used for classifying data including linear and nonlinear, which not only has a solid theoretical foundation, but also is more accurate than other algorithms in many areas of applications, especially in dealing with high-dimensional data. In SVM, kernel function is an important component, which makes it not necessary for us to get the specific mapping function in solving quadratic optimization problem of SVM. The only thing we need to do is to use kernel function to replace the complicated calculation of the dot product of data set, which noticeably reduces the number of dimension calculation. In this paper, we will introduce the theoretical basis of support vector machine, summarize the research status and analyze the research direction and development prospects of kernel function.