Uniform Design method to Finding the Optimal Parameters for Support Vector Machine

Support vector machine (SVM) has become one of the most popular learning machines. However, it is difficult to constructing a highly effective model for predicting before the parameter of SVM are optimally determined. Two parameters, C and g, must be determined in the establishment of an efficient SVM model. The purpose of this paper is to use uniform design method for finding the optimal values of the two parameters in establishing an efficient SVM model. According to the uniform design method, the regression relation between input parameters and output root-mean-square error can be obtained. Then, we can find the optimal parameters form the regression equation. The results prove that the proposed method can effectively improve the prediction accuracy.