Computational aspects of Bayesian spectral density estimation
Liseo, Brunero; Rousseau, Judith; Chopin, Nicolas (2013), Computational aspects of Bayesian spectral density estimation, Journal of Computational and Graphical Statistics, 22, 3, p. 533-557. http://dx.doi.org/10.1080/10618600.2013.785293
Type
Article accepté pour publication ou publiéExternal document link
http://hal.archives-ouvertes.fr/hal-00767466Date
2013Journal name
Journal of Computational and Graphical StatisticsVolume
22Number
3Publisher
Taylor and Francis; Taylor and Francis
Pages
533-557
Publication identifier
Metadata
Show full item recordAbstract (EN)
Gaussian time-series models are often specified through their spectral density. Such models pose several computational challenges, in particular because of the non-sparse nature of the covariance matrix. We derive a fast approximation of the likelihood for such models. We use importance sampling to correct for the approximation error. We show that the variance of the importance sampling weights vanishes as the sample size goes to infinity. We show that the posterior is typically multi-modal, and derive a Sequential Monte Carlo sampler based on an annealing sequence in order to sample from the approximate posterior. Performance of the overall approach is evaluated on simulated and real datasets.Subjects / Keywords
Sequential Monte Carlo; Long memory processes; FEXPRelated items
Showing items related by title and author.
-
Rousseau, Judith; Chopin, Nicolas; Liseo, Brunero (2012) Article accepté pour publication ou publié
-
Liseo, Brunero; Rousseau, Judith (2006) Document de travail / Working paper
-
Rousseau, Judith; Kruijer, Willem (2011) Document de travail / Working paper
-
Rousseau, Judith; Chopin, Nicolas; Cottet, Vincent; Alquier, Pierre (2014) Document de travail / Working paper
-
Kruijer, Willem; Rousseau, Judith; Van Der Vaart, Aad (2009-05) Communication / Conférence