基于MATLAB的组合模型

在港口吞吐量预测中的应用

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。-A combination of MATLAB-based model in port throughput prediction of a Hong Kong 1998- 2007 container throughput for the measured data, the establishment of its container throughput of the gray theory, the three exponential smoothing, prediction models, such as cubic polynomial in MATLAB, the contrast of the Hong Kong container throughput of fitting the models to predict the value of the difference between the actual value. Analysis of the causes of differences and the limitations of a single prediction model, the combination of port container throughput prediction. Its prediction error was significantly lower than other single model. The use of combination forecasting model. Can reduce the error. Improve the prediction accuracy.