首页 | 本学科首页   官方微博 | 高级检索  
     


Optimal conditional ARIMA forecasts
Authors:  ctor M. Guerrero
Abstract:An optimal univariate forecast, based on historical and additional information about the future, is obtained in this paper. Its statistical properties, as well as some inferential procedures derived from it, are indicated. Two main situations are considered explicitly: (1) when the additional information imposes a constraint to be fulfilled exactly by the forecasts and (2) when the information is only a conjecture about the future values of the series or a forecast from an alternative model. Theoretical and empirical illustrations are provided, and a unification of the existing methods is also attempted.
Keywords:ARIMA models  Combination of forecasts  Minimum mean-square error  Prior information  Quadratic minimization  Time series
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号