指数平滑法中平滑系数的选择研究

Selection of Smoothing Coefficient via Exponential Smoothing Algorithm

在理论与实例分析的基础上,分析了用指数平滑法选择平滑系数的不足,提出了指数平滑法在修匀数据和用于预测时的选择思路及方法.平滑系数的取值大小对结果的准确性具有重要作用,当指数平滑法应用于修匀数据时,平滑系数应取得小一些;而当其应用于预测时,则应取得大一些,这样可有效地增加结果的准确性.

After an analysis on the disadvantages of Exponential Smoothing Algorithm （ESA） in the selection of smoothing coefficient from theoretical and empirical perspectives, the author has proposed a new method for selecting smoothing coefficient by ESA to smooth and forecast data. The value of smoothing coefficient has a significant impact on the accuracy of data smoothing and data forecasting. when the ESA is used to smooth data, the smoothing coefficient needs to be smaller. When it is used to forecast data, the smoothing coefficient needs to be greater, so as to improve the accuracy of data smoothing and data forecasting.