A Study on Port Container Throughput Prediction Based on Optimal Combined Forecasting Model in Shanghai Port

Abstract

Port throughput is an important indicator for port planning and operations. It is essential to improve the accuracy of throughput prediction. In this study, by using the dataset of Shanghai Port container throughputs in the period of 1995–2010, based on the framework of PMVF, the performance of traditional models such as cubic exponential smoothing, GM (1,1), and multiple regression analysis single models have been evaluated. Comparison with the observed dataset of port container throughput indicates that traditional models show their limitations respectively with unsatisfied predictions. Based on the advantages of traditional single models, the optimal combined forecasting model is proposed in this study by comparing different possible combinations. Comparison with observed data shows that the optimal combined model improves the accuracy of forecasting container throughput in Shanghai port. The further predictions on container throughput of Shanghai port from 2010 to 2011 can be served for the port's planning and operations.