A Hybrid ARIMA and SVM Model for Traffic Flow Prediction Based on Wavelet Denoising
Jianmin Xu · Journal of Highway and Transportation Research and Development · 2009
Based on the analysis of the characteristics of nonlinearity and strong interference of traffic flow due to the complex and uncertainty of time variance in real traffic system,a new approach was proposed for traffic flow prediction.First,wavelet transform is employed to eliminate the noise of original traffic data to reflect the essence and variation of traffic flow.Then a hybrid methodology that exploits the unique strength of the ARIMA model and the SVM model to forecast traffic flow with the worked data was proposed.Finally,numerical field examples were given to testify the precision of the model.The result shows that(1)the hybrid model can produce more accurate predictions than that of single model;(2)the hybrid model that uses the method of wavelet denoising is more efficient and reliable.The hybrid model based on wavelet denoising can be an efficient method to the real-time dynamic traffic flow prediction.