摘要
用组合预测法对区域物流需求进行预测,各单项预测方法组合前需进行有效性分析。定性分析后选择回归分析、灰色预测、指数平滑法,定量计算验证时以1991~2004年综合货运量的预测为例,求出三种单项预测方法的预测精度序列值及相对误差值,依据组合预测的有效性对计算结果进行比较分析,经计算本文最终选定回归分析、指数平滑和灰色模型作为非线性组合预测模型的输入方法。
If the demand of the regional logistics was forecasted by using the combination forecast, it was necessary to analyze the validity of the single item forecast method before they were combined. After qualitatively analyzing, the three kinds of method: regression analysis, the gray forecast, and the index flatness method was chosen to calculate quantitatively. It was validated by qualitative calculation, taking our country 1991 ~2004 year comprehensive transported goods volume forecast as an example. The author extracted the sequence value of the forecast precision and the relative error value of three single items forecasts method, and carried on the comparative analysis based on the combination forecast validity to the computed result. After computation in this paper, the regression analysis, the index flatness method, and the gray model, finally is selected as input method of the non-linear combination forecast model.
出处
《工业工程与管理》
2008年第1期94-98,共5页
Industrial Engineering and Management
基金
国家自然科学基金资助项目(70672102)
关键词
组合预测
预测的有效度
线性组合
非线性组合
combinatiion forecasts
effectiveness of reckoning
linear combination
non-linear combination