一次指数平滑法是指以最后的一个第一次指数平滑。如果为了使指数平滑值敏感地反映最新观察值的变化，应取较大阿尔法值，如果所求指数平滑值是用来代表该时间序列的长期趋势值，则应取较小阿尔法值。(An exponential smoothing method refers to smoothing with the last first exponential. If the exponent smoothing value sensitively reflects the change of the latest observed value, a larger alpha value should be taken. If the exponent smoothing value is used to represent the long-term trend value of the time series, a smaller alpha value should be taken.)