动态三次指数平滑法及应用实例

Dynamic Three Exponential Smoothing Method and Its Application

传统的指数模型多用于处理线性趋势的时间序列,本文在此基础上建立了可以处理非线性时间序列数据的动态三次指数平滑模型.以拟合值与源数据误差的平方和为评价指标,通过使得该指标最小来计算最优时间序列系数,建立三次动态指数平滑模型.评价指标类似于评价函数,对模型自动评价,从而提高模型的适应能力,提高模型的计算精度.通过对实际时间序列的分析,进一步证明通过误差评价指标评价的方法建立的动态指数平滑可以降低误差,使预测更加精确.

Traditional exponential model is usually used to handle the linear trend of time series. In this paper the model is built up based on the nonlinear dynamic three exponential smoothing model of time series data. To fit the square error value and the source of data for the evaluation and by making the minimum index to calculate the optimal time series, the three dynamic exponential smoothing model is set up. Evaluation index is similar to the evaluation function, the automatic evaluation of the model, so as to improve the adaptability of the model and the calculation accuracy of the models. Through the analysis of actual time series, it further proves that dynamic exponen- tial smoothing established through error evaluation index can reduce the error of the prediction so as to make it more accurate.