Application of Combined Model in Forecasting Logistic Volume of a Port

Firstly, this paper predicts the volume of logistics in Ningbo Port with the two methods of improved BP neural network and gray model, and introduces combined forecast methods on the basic of researching on those two forecast methods. Theories and practices have shown that combined forecast model are more accurate than single forecast model, and can enhance the stability of the forecast, thus own higher ability to predict environmental change. Based on the Shapley value allocation of the combined forecast model, this paper aims to study the demand forecast of the logistics. The two methods mentioned inferior are realized by matlab programming.