Scalable Monte Carlo inference for state-space models

Sinan Yıldırım, Christophe Andrieu, Arnaud Doucet

Research output: Contribution to journalArticle (Academic Journal)

33 Downloads (Pure)

Abstract

We present an original simulation-based method to estimate likelihood ratios efficiently for general state-space models. Our method relies on a novel use of the conditional Sequential Monte Carlo (cSMC) algorithm introduced in \citet{Andrieu_et_al_2010} and presents several practical advantages over standard approaches. The ratio is estimated using a unique source of randomness instead of estimating separately the two likelihood terms involved. Beyond the benefits in terms of variance reduction one may expect in general from this type of approach, an important point here is that the variance of this estimator decreases as the distance between the likelihood parameters decreases. We show how this can be exploited in the context of Monte Carlo Markov chain (MCMC) algorithms, leading to the development of a new class of exact-approximate MCMC methods to perform Bayesian static parameter inference in state-space models. We show through simulations that, in contrast to the Particle Marginal Metropolis-Hastings (PMMH) algorithm of Andrieu_et_al_2010, the computational effort required by this novel MCMC scheme scales very favourably for large data sets.
Original languageEnglish
Number of pages35
JournalarXiv
Publication statusSubmitted - 7 Sep 2018

Keywords

  • Annealed importance sampling
  • Particle Markov chain Monte Carlo
  • Sequential Monte Carlo
  • State-space models

Fingerprint Dive into the research topics of 'Scalable Monte Carlo inference for state-space models'. Together they form a unique fingerprint.

  • Cite this

    Yıldırım, S., Andrieu, C., & Doucet, A. (2018). Scalable Monte Carlo inference for state-space models. Manuscript submitted for publication.