On the Containment Condition for Adaptive Markov Chain Monte Carlo Algorithms
نویسندگان
چکیده
This paper considers ergodicity properties of certain adaptive Markov chain Monte Carlo (MCMC) algorithms for multidimensional target distributions, in particular Adaptive Metropolis and Adaptive Metropoliswithin-Gibbs. It was previously shown by Roberts and Rosenthal (2007) that Diminishing Adaptation and Containment imply ergodicity of adaptive MCMC. We derive various sufficient conditions to ensure Containment, and connect the convergence rates of algorithms with the tail properties of the corresponding target distributions. Two examples are given to show that Diminishing Adaptation alone does not imply ergodicity. We also present a Summable Adaptive Condition which, when satisfied, proves ergodicity more easily. One example is given to explain that summable adaptive condition is sufficient for Containment.
منابع مشابه
Convergence of Adaptive Markov Chain Monte Carlo Algorithms
In the thesis, we study ergodicity of adaptive Markov Chain Monte Carlo methods (MCMC) based on two conditions (Diminishing Adaptation and Containment which together imply ergodicity), explain the advantages of adaptive MCMC, and apply the theoretical result for some applications. First we show several facts: 1. Diminishing Adaptation alone may not guarantee ergodicity; 2. Containment is not ne...
متن کاملThe Containment Condition and Adapfail Algorithms
This short note investigates convergence of adaptive MCMC algorithms, i.e. algorithms which modify the Markov chain update probabilities on the fly. We focus on the Containment condition introduced in [RR07]. We show that if the Containment condition is not satisfied, then the algorithm will perform very poorly. Specifically, with positive probability, the adaptive algorithm will be asymptotica...
متن کاملSimultaneous drift conditions for Adaptive Markov Chain Monte Carlo algorithms
In the paper, we mainly study ergodicity of adaptive MCMC algorithms. Assume that under some regular conditions about target distributions, all the MCMC samplers in {Pγ : γ ∈ Y} simultaneously satisfy a group of drift conditions, and have the uniform small set C in the sense of the m-step transition such that each MCMC sampler converges to target at a polynomial rate. We say that the family {Pγ...
متن کاملSpatial count models on the number of unhealthy days in Tehran
Spatial count data is usually found in most sciences such as environmental science, meteorology, geology and medicine. Spatial generalized linear models based on poisson (poisson-lognormal spatial model) and binomial (binomial-logitnormal spatial model) distributions are often used to analyze discrete count data in which spatial correlation is observed. The likelihood function of these models i...
متن کاملA Markov Chain Perspective on Adaptive Monte Carlo Algorithms
This paper discusses some connections between adaptive Monte Carlo algorithms and general state space Markov chains. Adaptive algorithms are iterative methods in which previously generated samples are used to construct a more efficient sampling distribution at the current iteration. In this paper, we describe two such adaptive algorithms, one arising in a finite-horizon computation of expected ...
متن کامل