A hybrid approach of traffic volume forecasting based on wavelet transform, neural network and Markov model

Traffic volume forecasting is an essential component of any responsive traffic control or route guidance system. A new traffic volume prediction approach is proposed based on wavelet transform, neural network and Markov model. First, apply multi-resolution analysis that is decomposition and reconstruction to the original traffic volume time series to obtain a trend series and a hierarchy of detail series that are easy to model and predict. Then a neural network is trained to provide a prediction to this trend series, and a Markov model is established for each detail series and these Markov models are used to predict the detail series. The combination of all these forecasting values, i.e. a prediction of trend series and a hierarchy prediction of detail series provides a final prediction to the original traffic volume series. This method´s performances are validated by a real traffic volume time series obtained in Suzhou city.