Multiscale Wavelet Support Vector Regression for Traffic Flow Prediction

Traffic flow is a fundamental measure in transportation. Accurate traffic flow prediction also is crucial to the development of intelligent transportation systems and advanced traveler information systems. A novel multiscale wavelet support vector regression method (MW-SVR) is proposed for traffic flow prediction. Based on wavelet multi-resolution analysis, a scaling kernel function with multi-resolution characteristics is constructed, implements the combination of the wavelet technique with support vector regression. A variety of experiments are carried out. The experimental results demonstrate that the proposed approach with multiscale wavelet kernel provides more optimal performance than that with radial basis function kernel, and the feasibility of applying MW-SVR in traffic flow prediction.