Research on Freight Traffic Forecast Based on Wavelet and Support Vector Machine

Abstract

This paper first carries on the elaboration to the least squares support vector machine (LS-SVM) forecast model. Basing on the theory of wavelet frame and the condition of the SVM kernel function, a method that generates wavelet kernel function of the support vector machine is proposed. Then the Mexican Hat wavelet is been selected to construct LS-SVM kernel function and form LS-SVM model based on the wavelet kernel function (the LS-WSVM model), after that forecast freight traffic of highway by this model in China. Through the contrast of forecast result between four different kernel functions, it indicated that the model using wavelet kernel function have a higher validity than that of other kernel functions. At the same time, contrasting the results between LS-WSVM forecast and other forecast methods, it also indicated the LS-WSVM is able to increase the forecast precision. After taking the model into different areas of china, we find that the model has the higher application value.