- Aktuality
- Předmět a cíle výzkumu
- Rada Centra a výroční zprávy
- Pracoviště a kontakty
- Přednášky
- Konference a semináře
- Publikace
- Ukázky aplikací
- Zahraniční cesty
- Hosté centra
- Minulé události
- Vstup do Synergy
VYHLEDÁVÁNÍ
PRACOVIŠTĚ CENTRA
Výzkumné centrum
Data - Algoritmy - Rozhodování
Data - Algoritmy - Rozhodování
Založeno v roce 2005 s podporou MŠMT ČR (projekt 1M0572)
Publikace
Mixture-Based Extension of the AR Model and its Recursive Bayesian Identification
Typ:
Článek v odborném periodiku
Autoři publikace:
Šmídl V., Quinn A.
Název periodika:
IEEE Transactions on Signal Processing
Řada:
53
Rok:
2005
Číslo:
9
Strany:
3530-3542
Kontaktní osoba:
Klíčová slova:
Bayesian identification, mixtures, sufficient statistics,
Anotace:
An extension of the AutoRegressive (AR) model is studied, which allows transformations and distortions on the regressor to be handled. Many important signal processing problems are amenable to this Extended AR (i.e. EAR) model. It is shown that Bayesian
identification and prediction of the EAR model can be performed recursively, in common with the AR model itself. The EAR model
does, however, require that the transformation be known. When it is unknown, the associated transformation space is represented by a finite set of candidates. What follows is a Mixture-based EAR model, i.e. the MEAR model. An approximate identification algorithm for MEAR is developed, using a restricted Variational Bayes (VB) method. This restores the elegant recursive update of sufficient statistics. The MEAR model is applied to the robust identification of AR processes corrupted by outliers and burst noise respectively, and to click removal for speech.
identification and prediction of the EAR model can be performed recursively, in common with the AR model itself. The EAR model
does, however, require that the transformation be known. When it is unknown, the associated transformation space is represented by a finite set of candidates. What follows is a Mixture-based EAR model, i.e. the MEAR model. An approximate identification algorithm for MEAR is developed, using a restricted Variational Bayes (VB) method. This restores the elegant recursive update of sufficient statistics. The MEAR model is applied to the robust identification of AR processes corrupted by outliers and burst noise respectively, and to click removal for speech.