On the LP-convergence of a Girsanov theorem based particle filter

We analyze the Lp-convergence of a previously proposed Girsanov theorem based particle filter for discretely observed stochastic differential equation (SDE) models. We prove the convergence of the algorithm with the number of particles tending to infinity by requiring a moment condition and a step-wise initial condition boundedness for the stochastic exponential process giving the likelihood ratio of the SDEs. The practical implications of the condition are illustrated with an Ornstein-Uhlenbeck model and with a non-linear Benes model.

[1]  Simo Särkkä,et al.  Recursive Bayesian inference on stochastic differential equations , 2006 .

[2]  P. Moral.  Nonlinear filtering : Interacting particle resolution , 1997 .

[3]  Randal Douc,et al.  Nonlinear Time Series: Theory, Methods and Applications with R Examples , 2014 .

[4]  Florian Nadel,et al.  Stochastic Processes And Filtering Theory , 2016 .

[5]  Xiao-Li Hu,et al.  A General Convergence Result for Particle Filtering , 2011, IEEE Transactions on Signal Processing.

[6]  P. Moral,et al.  Branching and interacting particle systems. Approximations of Feynman-Kac formulae with applications to non-linear filtering , 2000 .

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

[8]  I. Mbalawata.  Adaptive Markov chain Monte Carlo and Bayesian filtering for state space models , 2014 .

[9]  Simo Särkkä,et al.  Moment conditions for convergence of particle filters with unbounded importance weights , 2014, Signal Process..

[10]  Haikady N. Nagaraja,et al.  Inference in Hidden Markov Models , 2006, Technometrics.

[11]  I. V. Girsanov.  On Transforming a Certain Class of Stochastic Processes by Absolutely Continuous Substitution of Measures , 1960 .

[12]  P. Hall,et al.  Martingale Limit Theory and Its Application , 1980 .

[13]  P. Hall,et al.  Martingale Limit Theory and its Application. , 1984 .

[14]  Frederick Daum.  Exact finite dimensional nonlinear filters for continuous time processes with discrete time measurements , 1984, The 23rd IEEE Conference on Decision and Control.

[15]  Jianqing Fan.  Nonlinear Time Series , 2003 .

[16]  P. Moral.  Feynman-Kac Formulae: Genealogical and Interacting Particle Systems with Applications , 2004 .

[17]  D. Crisan,et al.  Fundamentals of Stochastic Filtering , 2008 .

[18]  Simo Sarkka.  On Sequential Monte Carlo Sampling of Discretely Observed Stochastic Differential Equations , 2006, 2006 IEEE Nonlinear Statistical Signal Processing Workshop.

[19]  A. Doucet,et al.  A survey of convergence results on particle ltering for practitioners , 2002 .

[20]  Pierre Del Moral,et al.  Mean Field Simulation for Monte Carlo Integration , 2013 .

[21]  S. Sarkka,et al.  Application of Girsanov Theorem to Particle Filtering of Discretely Observed Continuous - Time Non-Linear Systems , 2007, 0705.1598.

[22]  Arnaud Doucet,et al.  A survey of convergence results on particle filtering methods for practitioners , 2002, IEEE Trans. Signal Process..

[23]  Vigirdas Mackevičius,et al.  The finiteness of moments of a stochastic exponential , 2003 .