| The rapid growth of China’s port throughput requires timely adjustment of port, and the accurate prediction of port throughput is particularly important to port’s development, transformation, resources etc. Since the financial crisis in2008, the growth trend of China’s port throughput has entered a new phase, port throughput growth slowed, many factors affect the throughput growth. Compared with the previous law, it has undergone major changes. Traditional prediction methods and models have been difficult to ensure the needs of the current forecasting work. In this paper, the object of study for the Chinese domestic coastal ports throughput, represented by Guangzhou port, is to analysis the influence factors of port throughput. And rough sets and genetic programming combined throughput prediction model was constructed to achieve accurate prediction of throughput.Firstly, this paper analyzes the factors affecting the throughput, and the establishment of throughput index system; Then, this paper uses an improved numerical attribute reduction algorithm based on neighborhood rough set to reduce properties which Important degree is little, obtain key indicators; finally, this paper uses genetic programming method to train samples of key indicators and build the throughput forecasting model.Finally, this paper uses the historical data of Guangzhou port in2000-2012to evaluate to the feasibility and effectiveness of the throughput prediction model. Then, the paper makes2013-2015throughput predictions and provides reference for Guangzhou port planning.Main achievements of this paper include:1, the establishment of the Guangzhou port as the representatives of the domestic coastal port throughput index system;2, using an improved numerical attribute reduction algorithm based on neighborhood rough set to reduce properties;3, the first time the rough set attribute reduction method used in combination with genetic programming port throughput predictions. |
PDF Full Text Request