Multi-Layer Support Vector Machine and its Application

Abstract

Based on Statistical Learning Theory (SLT), Support Vector Machine (SVM), which is a new kind of machine learning method that is used for classification and regression. SVM is considered as two layers learning machine since it maps the original space into a high dimensional feature space, i.e., input layer and high dimensional feature space layer. If the high dimensional feature space layer is considered as a new problem's input layer and the new problem is also solved by SVM, the new problem can be solved by SVMs named Multi-Layer SVM (MLSVM). MLSVM is composed of input layer and at least one layer high dimensional feature space layer. In this paper, m-th order ordinary differential equations are solved by MLSVM for regression. Experimental results indicate that MLSVM can effectively solve the problem of ordinary differential equations. Thus, MLSVM exhibits its great potential to solve other complex problems.