铁路货物周转量预测方法研究

Study of Forecast Methods for Railroad Freight Turnover Volume

应用灰数列预测与神经网络预测的组合模型对铁路货物周转量进行预测。选取不同的灰数列，得到灰预测值作为神经网络的输入；利用MATLAB拟合周转量实际值，并入神经网络的输入。建立2个组合模型，拟合值的加入提高了预测精度。

This paper forecasts railroad freight turnover volume with the combination of gray forecast model and nerve net- work forecast model. By selecting different gray lists the input of nerve network model is obtained. By using fitting turnover volume of MATLAB, puts its results into the nerve network input. Based on two models , the forecast precision is improved.