SMC^2: A sequential Monte Carlo algorithm with particle Markov chain Monte Carlo updates

We consider the generic problem of performing sequential Bayesian inference in a state-space model with observation process y, state process x and fixed parameter theta. An idealized approach would be to apply the iterated batch importance sampling (IBIS) algorithm of Chopin (2002). This is a sequential Monte Carlo algorithm in the theta-dimension, that samples values of theta, reweights iteratively these values using the likelihood increments p(y_t|y_1:t-1, theta), and rejuvenates the theta-particles through a resampling step and a MCMC update step. In state-space models these likelihood increments are intractable in most cases, but they may be unbiasedly estimated by a particle filter in the x-dimension, for any fixed theta. This motivates the SMC^2 algorithm proposed in this article: a sequential Monte Carlo algorithm, defined in the theta-dimension, which propagates and resamples many particle filters in the x-dimension. The filters in the x-dimension are an example of the random weight particle filter as in Fearnhead et al. (2010). On the other hand, the particle Markov chain Monte Carlo (PMCMC) framework developed in Andrieu et al. (2010) allows us to design appropriate MCMC rejuvenation steps. Thus, the theta-particles target the correct posterior distribution at each iteration t, despite the intractability of the likelihood increments. We explore the applicability of our algorithm in both sequential and non-sequential applications and consider various degrees of freedom, as for example increasing dynamically the number of x-particles. We contrast our approach to various competing methods, both conceptually and empirically through a detailed simulation study, included here and in a supplement, and based on particularly challenging examples.

[1]  N. Gordon,et al.  Novel approach to nonlinear/non-Gaussian Bayesian state estimation , 1993 .

[2]  N. Shephard,et al.  Stochastic Volatility: Likelihood Inference And Comparison With Arch Models , 1996 .

[3]  Jonathan A. Tawn,et al.  Statistics for Exceptional Athletics Records , 1995 .

[4]  N. Shephard,et al.  Estimation of an Asymmetric Stochastic Volatility Model for Asset Returns , 1996 .

[5]  Jun S. Liu,et al.  Sequential Monte Carlo methods for dynamic systems , 1997 .

[6]  G. Kitagawa.  A self-organizing state-space model , 1998 .

[7]  P. Moral,et al.  Central limit theorem for nonlinear filtering and interacting particle systems , 1999 .

[8]  Simon J. Godsill,et al.  On sequential Monte Carlo sampling methods for Bayesian filtering , 2000, Stat. Comput..

[9]  Hans Kiinsch,et al.  State Space and Hidden Markov Models , 2000 .

[10]  Radford M. Neal.  Annealed importance sampling , 1998, Stat. Comput..

[11]  W. Gilks,et al.  Following a moving target—Monte Carlo inference for dynamic Bayesian models , 2001 .

[12]  Nando de Freitas,et al.  Sequential Monte Carlo Methods in Practice , 2001, Statistics for Engineering and Information Science.

[13]  Geir Storvik,et al.  Particle filters for state-space models with the presence of unknown static parameters , 2002, IEEE Trans. Signal Process..

[14]  N. Shephard,et al.  Econometric analysis of realized volatility and its use in estimating stochastic volatility models , 2002 .

[15]  P. Moral,et al.  Sequential Monte Carlo samplers , 2002, cond-mat/0212648.

[16]  M. Steel,et al.  Inference With Non-Gaussian Ornstein-Uhlenbeck Processes for Stochastic Volatility , 2006 .

[17]  A. Gallant,et al.  Alternative models for stock price dynamics , 2003 .

[18]  Carlo Gaetan,et al.  Smoothing Sample Extremes with Dynamic Models , 2004 .

[19]  Pierre Del Moral,et al.  Feynman-Kac formulae , 2004 .

[20]  N. Chopin.  Central limit theorem for sequential Monte Carlo methods and its application to Bayesian inference , 2004, math/0508594.

[21]  O. Cappé,et al.  Population Monte Carlo , 2004 .

[22]  Simon M. Potter,et al.  Forecasting and Estimating Multiple Change-Point Models with an Unknown Number of Change Points , 2004 .

[23]  Eric Moulines,et al.  Inference in hidden Markov models , 2010, Springer series in statistics.

[24]  N. Oudjane,et al.  Stability and Uniform Particle Approximation of Nonlinear Filters in Case of Non Ergodic Signals , 2005 .

[25]  Gareth O. Roberts,et al.  A General Framework for the Parametrization of Hierarchical Models , 2007, 0708.3797.

[26]  Nicolas Chopin,et al.  Inference and model choice for sequentially ordered hidden Markov models , 2007 .

[27]  Ajay Jasra,et al.  On population-based simulation for static inference , 2007, Stat. Comput..

[28]  Arnaud Doucet,et al.  Particle methods for maximum likelihood estimation in latent variable models , 2008, Stat. Comput..

[29]  R. Douc,et al.  Limit theorems for weighted samples with applications to sequential Monte Carlo methods , 2008 .

[30]  Arnaud Doucet,et al.  Sequential Monte Carlo computation of the score and observed information matrix in state-space models with application to parameter estimation , 2009 .

[31]  Arnaud Doucet,et al.  An overview of sequential Monte Carlo methods for parameter estimation in general state-space models , 2009 .

[32]  P. Fearnhead,et al.  A sequential smoothing algorithm with linear computational cost. , 2010 .

[33]  G. Peters,et al.  Ecological non-linear state space model selection via adaptive particle Markov chain Monte Carlo (AdPMCMC) , 2010, 1005.2238.

[34]  A. Doucet,et al.  Particle Markov chain Monte Carlo methods , 2010 .

[35]  Nicholas G. Polson,et al.  Particle Learning and Smoothing , 2010, 1011.1098.