Research on Forecasting Model in Short Term Traffic Flow Based on Data Mining Technology

Abstract

Accurate forecasting in short term traffic flow is of vital importance in management of good road traffic. In order to increase the precision, this paper proposes a forecasting model in short term traffic flow based on data mining technology. The model consists of three stages: first, the rough set theory and the genetic algorithm are applied to select relevant forecasting variable to the traffic flow; second, training pattern of wavelet neural network which is similar to the forecast term is carried out by using data mining technology; finally the wavelet neural network is used to carry on forecasting the traffic flow. Through forecasting traffic flow at Xinhua Street in Huhehot, the result shows that this model has a higher precision and surpasses gray GM (1, 1) and the BP artificial neural network model, which provides a new reliable and effective way of forecasting short term traffic flow of nodes in urban road network