Forecasting Model of Short-Term Traffic Flow Based on Cloud Model

In order to improve the prediction precision of the short-term traffic flow, a prediction method of short-term traffic flow based on cloud model was proposed. The traffic flow was fit by cloud model. The history cloud and the present cloud were built by historical traffic flow and present traffic flow. The forecast cloud is produced by both clouds. Then, combining with the volume of the short-term traffic flow of an intersection in Guangzhou City, the model was calculated and simulated through programming. Max Absolute Error (MAE) and Mean Absolute percent Error (MAPE) were used to estimate the effect of prediction. The simulation results indicate that this prediction method is effective and advanced. The change of the historical and real time traffic flow is taken into account in this method. Because the short-term traffic flow is dealt with as a whole, the error of prediction is avoided. The prediction precision and real-time prediction are satisfied.