我国平板玻璃需求预测优选方法的探讨

Discussion on Preferred Method for Domestic Market Forecasting of Sheet Glass

阐述了需求预测的重要性,介绍了在平板玻璃需求预测中应用较多的移动平均、指数平滑、趋势外推等时间序列外推法,以及多元线性回归等方法,通过寻找适用的最优预测模型,对不同预测模型进行试算,对预测结果进行分析,得出在平板玻璃短期需求预测时,使用二元线性回归模型的结果更接近最优解这一结论。

Demand forecasting is not only the basic to management branch of government for holding right advancing orientation and setting up new policies but also the important indicator to enterprise for investment decision and designing marketing program or strategy. There are now a lot of demand forecasting methods or models ready to be used, but the time series extrapolation, such as running average, exponential smoothing,tendency extrapolation and multiple linear regression are used in sheet glass industry frequently. In order to forecast the future as accurately as possible, reduce forecast time greatly and obtain higher efficiency we attempt to find available optimization forecast models for multiplying effectiveness. From the try calculations and forecast result analysis we came to a conclusion that the result from the use of binary linear regression model can approach the optimum solution for short term demand forecast in sheet glass industry.