Short-term Traffic Flow Prediction Based on EMD-WTD-SVR

Traffic flow prediction plays an important role in route guidance and road network management and control. A short-term traffic flow prediction model based on empirical mode decomposition (EMD), wavelet threshold denoising (WTD) and support vector regression (SVR) is proposed in this paper. EMD algorithm and WTD algorithm are used to decompose and reduce the noise of time series, multiple prediction results are given by SVR prediction model, and the prediction data are obtained by superimposing the prediction results, so as to make the prediction data closer to the actual value. The test on the data of a section of M2 highway in Britain shows that the prediction error of the EMD-WTD-SVR model is significantly lower than that of the EMD-SVR, WTD-SVR and common baseline models, which demonstrates the feasibility of the decomposition and denoising model.