Papers by Rivo Rakotozafy
On applique deux méthodes de Monte Carlo à des modèles à espace d'état avec paramètres inconnus. ... more On applique deux méthodes de Monte Carlo à des modèles à espace d'état avec paramètres inconnus. La première est un algorithme de Monte Carlo par chaîne de Markov. La seconde est le filtrage particulaire. Nous comparons ces méthodes sur un modèle d'évolution de la biomasse dans une pêcherie.
The application of the Markov chain to modeling agricultural succession is well known. In most ca... more The application of the Markov chain to modeling agricultural succession is well known. In most cases, the main problem is the inference of the model, i.e. the estimation of the transition matrix. In this work we present methods to estimate the transition matrix from historical observations. In addition to the estimator of maximum likelihood (MLE), we also consider the Bayes estimator associated with the Jeffreys prior. This Bayes estimator will be approximated by a Markov chain Monte Carlo (MCMC) method. We also propose a method based on the sojourn time to test the adequation of Markov chain model to the dataset.
Mathematics and Computers in Simulation, 2009
In many situations it is important to be able to propose N independent realizations of a given di... more In many situations it is important to be able to propose N independent realizations of a given distribution law. We propose a strategy for making N parallel Monte Carlo Markov chains (MCMC) interact in order to get an approximation of an independent N-sample of a given target law. In this method each individual chain proposes candidates for all other chains. We prove that the set of interacting chains is itself a MCMC method for the product of N target measures. Compared to independent parallel chains this method is more time consuming, but we show through examples that it possesses many advantages. This approach is applied to a biomass evolution model.
In many situations it is important to be able to propose N independent realizations of a given di... more In many situations it is important to be able to propose N independent realizations of a given distribution law. We propose a strategy for making N parallel Monte Carlo Markov Chains (MCMC) interact in order to get an approximation of an independent N -sample of a given target law. In this method each individual chain proposes candidates for all other chains. We prove that the set of interacting chains is itself a MCMC method for the product of N target measures. Compared to independent parallel chains this method is more time consuming, but we show through examples that it possesses many advantages. This approach is applied to a biomass evolution model.
Uploads
Papers by Rivo Rakotozafy