Forecasting freight volume scientifically is very important to the decision making of transportation and economic development strategies. In the forecast of freight volume, it is hard to obtain the satisfactory results with the traditional individual econometric forecasting methods and models for its various relative factors. Compared with individual forecast methods, the combination forecast method can improve the precision and stability of forecast results, and it is widely used in practice. In order to achieve a more accurate forecast result for freight volume, after comparing the current methods and models for forecast and summarizing the basic idea of combined forecast, the paper provides the combination forecast model based on the RBF neural network. A fitting and forecast model based on the historical data of freight volume with various methods separately is established, and then the fitting and forecast results are mixed into the combination model for forecast by the RBF neural network. At last, a forecast example of freight volume in Fujian Province is presented, and the results prove that the model for forecasting freight volume is effective and feasible, and it has a good practical value.
展开▼