基于模型组合的铁路集装箱运量预测

根据铁路集装箱运量预测受到多因素影响以及非线性的特点，本文采用灰色关联分析法选取了影响集装箱运量的主要因素，提出了一种基于非线性灰色模型和神经网络模型组合的铁路集装箱运量预测方法. 该方法将非线性灰色预测模型的预测值作为输入，相应的实际集装箱货运量作为输出，建立了神经网络模型结构，并提出了相应的算法. 最后以实例分析了该模型的可行性和科学性. 实例分析表明：非线性灰色模型预测的最大误差为10.52％，而组合模型的预测误差最大为8.72％，说明文中提出的组合预测模型充分考虑了多指标的共同作用，灰色预测模型提供了较完善的输入数据，神经网络模型考虑了各主要指标的关联关系.

Freight Volume Forecasting of Railway Container Based on Model Combination

With the characteristic of multiple factor and nolinear on freight volume forecasting of railway container, primary influencing factor of container freight volume are selected with gray association analysis in this paper. Freight volume forecasting of railway container model is put forward based on nolinear gray model and neural net model. The results of nolinear gray model are used as inputs and the actual freight volume is used as outputs. Then, a neural network was built and corresponding algorithm is designed. Finally, a case study is carried out to prove the validity, objectivity, and applicability of this model through calculated and analyzed practical data. From the test results, it is shown that maximum error of nolinear gray model is 10.52%, and maximum error of combination forecasting model is 8.72%. The nolinear gray forecasting model provides good input data sequences, and the relationships of indices are involved in the neural network.