We develop a Markov Chain Monte Carlo (MCMC) version of the Fourier domain Correlated Component Analysis (CCA) method. The MCMC version allow us to calculate the estimation error. However, an adaptive temperature and random walk step size parameter update rules are proposed for Metropolis scheme. The algorithm tested on astrophysical component separation problem.
MCMC correlated component analysis in Fourier domain
Salerno E;Kuruoglu E E
2009
Abstract
We develop a Markov Chain Monte Carlo (MCMC) version of the Fourier domain Correlated Component Analysis (CCA) method. The MCMC version allow us to calculate the estimation error. However, an adaptive temperature and random walk step size parameter update rules are proposed for Metropolis scheme. The algorithm tested on astrophysical component separation problem.File in questo prodotto:
| File | Dimensione | Formato | |
|---|---|---|---|
|
prod_161099-doc_133646.pdf
solo utenti autorizzati
Descrizione: MCMC Correlated Component Analysis in Fourier Domain
Dimensione
209.28 kB
Formato
Adobe PDF
|
209.28 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
