Long-term stability of sequential Monte Carlo methods under verifiable conditions

DOI10.1214/13-AAP962zbMATH Open1429.62364arXiv1203.6898OpenAlexW3101877509MaRDI QIDQ744372FDOQ744372

Jimmy Olsson, Eric Moulines, Randal Douc

Publication date: 25 September 2014

Published in: The Annals of Applied Probability (Search for Journal in Brave)

Abstract: This paper discusses particle filtering in general hidden Markov models (HMMs) and presents novel theoretical results on the long-term stability of bootstrap-type particle filters. More specifically, we establish that the asymptotic variance of the Monte Carlo estimates produced by the bootstrap filter is uniformly bounded in time. On the contrary to most previous results of this type, which in general presuppose that the state space of the hidden state process is compact (an assumption that is rarely satisfied in practice), our very mild assumptions are satisfied for a large class of HMMs with possibly noncompact state space. In addition, we derive a similar time uniform bound on the asymptotic $mathsf{L}^p$ error. Importantly, our results hold for misspecified models; that is, we do not at all assume that the data entering into the particle filter originate from the model governing the dynamics of the particles or not even from an HMM.

Full work available at URL: https://arxiv.org/abs/1203.6898

Recommendations

• Stability properties of some particle filters
• Sequential Monte Carlo Samplers: Error Bounds and Insensitivity to Initial Conditions
• A general theory of particle filters in hidden Markov models and some applications
• Uniform Stability of a Particle Approximation of the Optimal Filter Derivative
• Stability and uniform approximation of nonlinear filters using the Hilbert metric and application to particle filters

zbMATH Keywords

asymptotic variancesequential Monte Carlo methodsgeneral hidden Markov modelsbootstrap particle filterlocal Doeblin conditiontime uniform convergence

Mathematics Subject Classification ID

Asymptotic properties of parametric estimators (62F12) Markov processes: estimation; hidden Markov models (62M05) Inference from stochastic processes and prediction (62M20) Signal detection and filtering (aspects of stochastic processes) (60G35)

Cites Work

• Sequential Monte Carlo Methods in Practice
• Approximation Theorems of Mathematical Statistics
• Title not available (Why is that?)
• Stability and uniform approximation of nonlinear filters using the Hilbert metric and application to particle filters
• Inference in hidden Markov models.
• Filtering via Simulation: Auxiliary Particle Filters
• Title not available (Why is that?)
• On the stability of interacting processes with applications to filtering and genetic algorithms
• A sequential particle filter method for static models
• Limit theorems for weighted samples with applications to sequential Monte Carlo methods
• Stability and Uniform Particle Approximation of Nonlinear Filters in Case of Non Ergodic Signals
• Monte Carlo Smoothing for Nonlinear Time Series
• Sequential Monte Carlo smoothing for general state space hidden Markov models
• Recursive Monte Carlo filters: algorithms and theoretical analysis
• Fundamentals of stochastic filtering
• Title not available (Why is that?)
• Subadditive ergodic theory
• On the auxiliary particle filter
• Central limit theorem for nonlinear filtering and interacting particle systems
• Title not available (Why is that?)
• The Monte-Carlo method for filtering with discrete-time observations
• Forgetting the initial distribution for hidden Markov models
• A robustification approach to stability and to uniform particle approximation of nonlinear filters: the example of pseudo-mixing signals.
• Uniform approximations of discrete-time filters
• On discrete time ergodic filters with wrong initial data
• The stability of conditional Markov processes and Markov chains in random environments
• A note on auxiliary particle filters
• Exponential forgetting and geometric ergodicity for optimal filtering in general state-space models
• Stability of the discrete time filter in terms of the tails of noise distributions
• Asymptotic properties of the maximum likelihood estimation in misspecified hidden Markov models
• Moderate deviations for particle filtering
• Asymptotic properties of particle filter-based maximum likelihood estimators for state space models
• Convergence of empirical processes for interacting particle systems with applications to nonlinear filtering

Cited In (15)

• Stability with respect to initial conditions in V-norm for nonlinear filters with ergodic observations
• Stability of sequential Monte Carlo samplers via the Foster-Lyapunov condition
• Twisted particle filters
• Piecewise deterministic Markov processes for continuous-time Monte Carlo
• On the convergence of adaptive sequential Monte Carlo methods
• On the mathematical theory of ensemble (linear-Gaussian) Kalman-Bucy filtering
• Fluctuations, stability and instability of a distributed particle filter with local exchange
• On the stability of positive semigroups
• A sharp first order analysis of Feynman-Kac particle models. I: Propagation of chaos
• Path storage in the particle filter
• Error estimates on ergodic properties of discretized Feynman-Kac semigroups
• Limit theorems for sequential MCMC methods
• Particle-based online estimation of tangent filters with application to parameter estimation in nonlinear state-space models
• Numerically stable online estimation of variance in particle filters
• On the Behaviour of the Backward Interpretation of Feynman-Kac Formulae Under Verifiable Conditions

This page was built for publication: Long-term stability of sequential Monte Carlo methods under verifiable conditions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q744372)