Estimation and prediction with ARMMAX model: a mixture of ARMAX models with common ARX part

0411070 - UTIA-B 20030057 RIV GB eng J - Článek v odborném periodiku
He, L. - Kárný, Miroslav
Estimation and prediction with ARMMAX model: a mixture of ARMAX models with common ARX part.
International Journal of Adaptive Control and Signal Processing. Roč. 17, č. 4 (2003), s. 265-283. ISSN 0890-6327. E-ISSN 1099-1115
Grant CEP: GA AV ČR IBS1075102
Výzkumný záměr: CEZ:AV0Z1075907
Klíčová slova: Bayesian estimation * ARMAX model * finite mixtures
Kód oboru RIV: BD - Teorie informace
Impakt faktor: 0.602, rok: 2003
http://library.utia.cas.cz/separaty/historie/karny-estimation and prediction with armmax model a mixture of armax models with common arx part.pdf

The paper concerns Bayesian estimation and prediction in the presence of colored noise. It introduces a novel model called ARMMAX, a finite mixture with its ARMAX components having a common ARX part. A modified recursive Quasi-Bayes algorithm is proposed for its estimation, based on a classical solution that allows the MA part to be even at the stability boundary. An off-line use of the mixture offers the possibility to estimate C-parameters of an ARMAX model in a novel way.
Trvalý link: http://hdl.handle.net/11104/0131157