Short-Term Traffic Flow Combined Forecasting Model Based on SVM

This research concerns itself with a wavelet-SVM combined model study of the short-term traffic flow prediction issue. Different theories and methods have been introduced in the field to solve short-term traffic flow forecasting problem. And in our study, we attempt to use an alternative prediction framework to examine the combined model. This paper consists of four sections. A brief introduction is given in Section one of this study. Section two includes the theories of wavelet and support vector machine (SVM), then put forward the combined model. Section three focuses on a numerical study based on the actual speed data of an expressway in Beijing The whole paper ends with the conclusion that the combined model has very high accuracy.