平滑系数自适应的二次指数平滑模型及其应用

通过对传统指数平滑模型的分析,提出了动态平滑参数的概念;并由此建立了平滑权重对时间序列能够自适应的新的二次指数平滑新模型;进而得到Brown单参数和Holt双参数两类线性趋势模型及其不同于传统模型的良好性质.新模型使困绕指数平滑应用的初值难以选取、平滑参数适应性差及系统预测偏差等问题得到了较完整的解决.预测实例表明新模型提高了预测精度,并有很好的自适应性.

Quadric Exponential Smoothing Model with Adapted Parameter and Its Applications

After analyzing traditional quadratic exponential smoothing model,this paper puts forward the concept of dynamic smoothing parameter and sets sup the new model which is automaticaly adapted to the giwen time series?Furthermore,we deduce two kinds of linear trend models with Brown single parameter and Holt double parameter, and the good characters which is different with the traditional model?So some questions, i.e., the smoothing parameter is static and determined by one's experiences, and the smoothing ini...