Research on Container Throughput Forecast of Qingdao Port Based on Combined Forecasting Model

Abstract

The container throughput reflects the development level of the port to a large extent, and the scientific prediction of the container throughput of a port is conducive to the reasonable planning and construction of the port. Based on the historical data of container throughput of Qingdao Port, this paper establishes three kinds of single prediction models, namely, three-order exponential smoothing model, logistic growth curve function model and BP neural network model, and combined forecasting models. By comparing the average relative errors of the fitted values of the four models, it can be proved that the combined forecasting model enjoys more precise forecast than monomial forecast model. Finally, according to the prediction results of the combined model, the corresponding development strategy is proposed for Qingdao Port.