残缺数据的指数平滑模型

Smoothing Methods for Time Series With Missing Observations

在实际的预测工作中,搜集到的数据,通常不是等时间间距的而是残缺的,怎样利用这种不完整的数据,是预测工作中的难题。本文分析了残缺数据对预测精度的影响并就如何有效地利用MAGNE模型进行了论述。

Frequently a time series contains missing observations in stead of equally space observations. So, how to use the information is a difficult problem in forcasting. This paper analyses the missing observations' influence on the results of forcasting. It presents two modified methods to handle missing data within the smoothing methods.