Modeling Average Daily Traffic Volume using Neural Network-Wavelet Hybrid Method

Abstract

Forecasting traffic volume accurately and in a timely manner plays an important role to providing real-time traffic information, reducing congestion in pathways, and improving traffic safety. A combination of multi-layer back-propagation neural networks (BPNN) and wavelet transform is used for forecasting average daily traffic volume. Real data used in modeling are taken from the Qom-Tehran road during 2006-2008. Given the proposed method (WBPNN), the traffic volume data were initially preprocessed using wavelet transform. The input signal (the daily traffic volume time series) is decomposed into low- and high-frequency components up to 5 levels using the mother wavelet function Haar, so that more complete information would be obtained regarding the problem dynamics. The processed data are then fed to the neural network as training and test data. The trained network is validated considering evaluation functions such as MAE, MAPE, and VAPE. The results indicate that the proposed method predicts daily traffic volume with great precision and puts forward a model using native parameters, in addition to increased prediction accuracy.