Fluctuations of Interacting Markov Chain Monte Carlo Models

نویسندگان

  • Bernard Bercu
  • Arnaud Doucet
  • Pierre Del Moral
چکیده

We present a functional central limit theorem for a general class of interacting Markov chain Monte Carlo interpretations of discrete generation measure-valued equations. The path space models associated with these stochastic processes belong to the class of nonlinear Markov chains interacting with their empirical occupation measures. We develop an original theoretical analysis based on resolvent operators and semigroup techniques to analyze the fluctuation of their occupation measures around their limiting value. We also present a set of simple regularity conditions that applies to interacting Markov chain Monte Carlo models on path spaces, yielding what seems to be the first fluctuation theorems for this class of self-interacting models. Key-words: Multivariate and functional central limit theorems, random fields, martingale limit theorems, self-interacting Markov chains, Markov chain Monte Carlo models. ∗ Centre INRIA Bordeaux et Sud-Ouest & Institut de Mathématiques de Bordeaux , Université de Bordeaux I, 351 cours de la Libération 33405 Talence cedex, France, [email protected] † Centre INRIA Bordeaux et Sud-Ouest & Institut de Mathématiques de Bordeaux , Université de Bordeaux I, 351 cours de la Libération 33405 Talence cedex, France, [email protected] ‡ Department of Statistics, University of British Columbia, 333-6356 Agricultural Road, Vancouver, BC, V6T 1Z2, Canada, [email protected] in ria -0 02 27 53 6, v er si on 5 5 Fe b 20 08 in ria -0 02 27 53 6, v er si on 5 5 Fe b 20 08 Fluctuations de Modèles de Monte Carlo par Châınes de Markov en Interaction Résumé : Nous présentons un théorème de la limite centrale fonctionnel pour une classe générale d’algorithmes de Monte Carlo par châınes de Markov en interaction, utilisés dans la résolution numérique d’équations à valeurs mesures non linéaires. Les modèles trajectoriels associés à ces processus stochastiques appartiennent à la classe des modèles de châınes de Markov non linéaires, en interaction avec leurs mesures d’occupation temporelle. Nous développons une analyse des fluctuations originale fondée sur l’étude fine d’opérateurs résolvants et sur des techniques de semigroupes sur des espaces de distributions. Cette étude dépend d’un jeu de conditions de régularité simples permettant d’analyser des modèles de Monte Carlo par châınes de Markov en interaction sur des espaces trajectoriels. Ces résultats semblent être les premiers de ce type pour ces classes de processus en auto-interaction. Mots-clés : Théorèmes de la limite centrale multidimensionnels, champs aléatoires, théorèmes limites de martingales, processus en auto-interaction, méthodes de Monte Carlo par châınes de Markov in ria -0 02 27 53 6, v er si on 5 5 Fe b 20 08 in ria -0 02 27 53 6, v er si on 5 5 Fe b 20 08 Fluctuations of Interacting Markov Chain Monte Carlo Models 3

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Fluctuations of Interacting Markov Chain Monte Carlo Methods

We present a multivariate central limit theorem for a general class of interacting Markov chain Monte Carlo algorithms used to solve nonlinear measure-valued equations. These algorithms generate stochastic processes which belong to the class of nonlinear Markov chains interacting with their empirical occupation measures. We develop an original theoretical analysis based on resolvent operators a...

متن کامل

A Functional Central Limit Theorem for a Class of Interacting Markov Chain Monte Carlo Models

We present a functional central limit theorem for a new class of interaction Markov chain Monte Carlo interpretations of discrete generation measure valued equations. We provide an original stochastic analysis based on semigroup techniques on distribution spaces and fluctuation theorems for selfinteraction random fields. Besides the fluctuation analysis of these models, we also present a series...

متن کامل

Spatial 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 کامل

A Functional Central Limit Theorem for a Class of Interacting Markov Chain Monte Carlo Methods

We present a functional central limit theorem for a new class of interacting Markov chain Monte Carlo algorithms. These stochastic algorithms have been recently introduced to solve non-linear measure-valued equations. We provide an original theoretical analysis based on semigroup techniques on distribution spaces and fluctuation theorems for self-interacting random fields. Additionally we also ...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2008