Robust methods in exponential smoothing of time series. Abstract

0410886 - UTIA-B 20020100 SK eng A - Abstrakt
Michálek, Jiří
Robust methods in exponential smoothing of time series. Abstract.
Bratislava: Faculty of Mathematics, 2002. Abstracts of the Fourth International Conference on Mathematical Statistics. ProbaStat 2002. - (Witkovský, V.). s. 32-34
[ProbaStat 2002 /4./. 04.02.2002-08.02.2002, Smolenice]
Grant CEP: GA ČR GA402/01/1548
Výzkumný záměr: CEZ:AV0Z1075907
Klíčová slova: robust statistics * exponential smoothing
Kód oboru RIV: BB - Aplikovaná statistika, operační výzkum

Exponential smoothing is a very famous and often used recursive procedure for both smoothing and prediction of time series. This procedure is relatively very simple and gives good results in practice. However, in case when data are corrupted by outliers, the classical technique of exponential smoothing can fail. This situation calls for an application of robust statistics methods to exponential smoothing.
Trvalý link: http://hdl.handle.net/11104/0130973