港口集装箱吞吐量的灰色神经网络预测模型研究

为了降低港口集装箱吞吐量的预测误差，提高预测精度，文章通过分析传统的灰色预测模型和 BP 神经网络预测模型的优缺点，构建了灰色神经网络港口集装箱吞吐量预测模型，该模型充分发挥了灰色模型所需初始数据少和 BP 神经网络非线性拟合能力强的特点。以实际数值作为初始数据，各种灰色模型的预测值为神经网络的输入值，神经网络的输出值为组合预测结果。通过实例分析，结果表明：灰色神经网络预测模型提高了预测精度，预测结果比较理想，优于单一预测模型，因此，该模型用于港口集装箱吞吐量预测是可行的、有效的。

Prediction model of port container throughput with grey neural network

In order to reduce prediction error of port container throughput and improve its prediction accuracy, the grey neural network model of port container throughput is constructed after the advantages and disadvantages of the conventional grey model and BP neural network model have been analyzed. The new model gives full play to the characters of low data demand of grey model and strong nonlinear fitting ability of BP neural network. It u-ses actual measured values as the initial data,various prediction values of grey model as input data of neural net-work and final output data of neural network as combined prediction result. A case study shows that,better than a single forecasting model,the grey neural network model can offer improved prediction accuracy and ideal pre-diction result. Therefore,it is feasible and effective to use the model predict port container throughput.