Particle Markov Chain Monte Carlo for Efficient Numerical Simulation
نویسندگان
چکیده
Markov Chain Monte Carlo (MCMC) and sequential Monte Carlo (SMC) methods are the two most popular classes of algorithms used to sample from general high-dimensional probability distributions. The theoretical convergence of MCMC algorithms is ensured under weak assumptions, but their practical performance is notoriously unsatisfactory when the proposal distributions used to explore the space are poorly chosen and/or if highly correlated variables are updated independently. We show here how it is possible to systematically design potentially very efficient high-dimensional proposal distributions for MCMC by using SMC techniques. We demonstrate how this novel approach allows us to design effective MCMC algorithms in complex scenarios. This is illustrated by a problem of Bayesian inference for a stochastic kinetic model.
منابع مشابه
Markov chain Monte Carlo methods for visual tracking
Tracking articulated figures in high dimensional state spaces require tractable methods for inferring posterior distributions of joint locations, angles, and occlusion parameters. Markov chain Monte Carlo (MCMC) methods are efficient sampling procedures for approximating probability distributions. We apply MCMC to the domain of people tracking and investigate a general framework for sample-appr...
متن کاملMonte Carlo Simulation to Solve the Linear Volterra Integral Equations of The Second Kind
This paper is intended to provide a numerical algorithm based on random sampling for solving the linear Volterra integral equations of the second kind. This method is a Monte Carlo (MC) method based on the simulation of a continuous Markov chain. To illustrate the usefulness of this technique we apply it to a test problem. Numerical results are performed in order to show the efficiency and accu...
متن کاملVolatility, Jumps and Predictability of Returns: a Sequential Analysis
In this paper we propose a sequential Monte Carlo algorithm to estimate a stochastic volatility model with leverage effects and non constant conditional mean and jumps. We are interested in estimating the time invariant parameters and the non-observable dynamics involved in the model. Our idea relies on the auxiliary particle filter algorithm mixed together with Markov Chain Monte Carlo (MCMC) ...
متن کاملBayesian parameter estimation in dynamic population model via particle Markov chain Monte Carlo
In nature, population dynamics are subject to multiple sources of stochasticity. State-space models (SSMs) provide an ideal framework for incorporating both environmental noises and measurement errors into dynamic population models. In this paper, we present a recently developed method, Particle Markov Chain Monte Carlo (Particle MCMC), for parameter estimation in nonlinear SSMs. We use one eff...
متن کاملOn Finding the Smallest Generalized Eigenpair Using Markov Chain Monte Carlo Algorithm
This paper proposes a new technique based on inverse Markov chain Monte Carlo algorithm for finding the smallest generalized eigenpair of the large scale matrices. Some numerical examples show that the proposed method is efficient.
متن کامل