交通运输工程与信息学报

铁路货运量预测是铁路运输市场分析的重要内容,对铁路货运业务的开展有着重要指导意义。铁路月度货运量数据序列是既有趋势性增长又有季节性波动特征的时间序列,Holt-Winters模型适用这类时间序列的预测。本文构建Holt-Winters乘法模型来进行铁路月度货运量预测,并以某铁路局化肥月度货运量为原始数据来做实证分析,最后与灰色模型、分组回归等传统预测模型的结果进行比较,结果显示Holt-Winters模型的预测精度最高。

Railway freight forecast is an important part of the analysis of railway transportation market, which is of great significance to the development of railway freight business. The railway monthly freight data sequence is such a time series with both trendy growth and seasonal fluctuating characteristics. Holt-Winters model is suitable to predict the time series. In this paper, a Holt-Winters multiplication model was constructed to forecast the railway monthly freight, and the railway bureau fertilizer railway monthly freight as original data was used to do an empirical analysis. The results were compared with the traditional forecasting models', such as gray model's and group regression's. The result shows that Holt-Winters model has the highest prediction accuracy.