基于MATLAB的组合模型在港口吞吐量预测中的应用  

Application of combined forecasting model based on MATLAB for forecasting of port throughput capacity

作　　者：刘启文[1] 林钢[1] 林吾思[2] 

机构地区：[1]中山大学,广东广州510275 [2]北京邮电大学,北京100876

出　　处：《水运工程》2008年第11期86-89,共4页Port & Waterway Engineering

摘　　要：以某港1998—2007年集装箱吞吐量为实测数据,建立其集装箱吞吐量的灰色理论、三次指数平滑、三次多项式等预测模型,在MATLAB下,对比该港集装箱吞吐量各模型预测拟合值与实际值的差异,分析了差异产生的原因及其单一预测模型的局限性,提出了港口集装箱吞吐量组合预测法。其预测误差明显低于其它单模型。运用组合预测模型,可以降低误差,提高预测精度。Taking the container throughput of a port during 1998-2007 as the original data, some prediction models of container throughput in this port are established by means of gray theory model, cubic flatness model, and cubic polynomial model. Comparing with the differences between the prediction value in each prediction model and the real value of this port＇s container throughput with MATLAB, the reasons of these differences and the limit of each single prediction model are analyzed. Then combined forecasting method is put forward to forecast port＇s container throughput and it＇s forecasting error is less than the others. So combined forecasting method is a better one for port＇s throughput forecasting and it can decrease the forecasting error and increase the forecasting precision.

关 键 词：MATLAB 集装箱吞吐量 组合预测模型 

分 类 号：U652.14[交通运输工程—港口、海岸及近海工程]