Convergence of adaptive mixtures of importance sampling schemes - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Statistics and Computing Année : 2008

Convergence of adaptive mixtures of importance sampling schemes

Résumé

In the design of efficient simulation algorithms, one is often beset with a poor choice of proposal distributions. Although the performance of a given simulation kernel can clarify a posteriori how adequate this kernel is for the problem at hand, a permanent on-line modification of kernels causes concerns about the validity of the resulting algorithm. While the issue is most often intractable for MCMC algorithms, the equivalent version for importance sampling algorithms can be validated quite precisely. We derive sufficient convergence conditions for adaptive mixtures of population Monte Carlo algorithms and show that Rao--Blackwellized versions asymptotically achieve an optimum in terms of a Kullback divergence criterion, while more rudimentary versions do not benefit from repeated updating.

Dates et versions

hal-00432955 , version 1 (17-11-2009)

Identifiants

Citer

Randal Douc, A. Guillin, Jean-Michel Marin, C. P. Robert. Convergence of adaptive mixtures of importance sampling schemes. Statistics and Computing, 2008, 18,, pp.447-459. ⟨hal-00432955⟩
279 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More