Volume 8, August 2004
|Page(s)||115 - 131|
|Published online||15 September 2004|
Coupling a stochastic approximation version of EM with an MCMC procedure
Université Paris Sud, Bât. 425, 91400 Orsay, France; Estelle.Kuhn@math.u-psud.fr.
2 Université René Descartes and Université Paris Sud, Bât. 425, 91400 Orsay, France; Marc.Lavielle@math.u-psud.fr.
Revised: 2 November 2003
The stochastic approximation version of EM (SAEM) proposed by Delyon et al. (1999) is a powerful alternative to EM when the E-step is intractable. Convergence of SAEM toward a maximum of the observed likelihood is established when the unobserved data are simulated at each iteration under the conditional distribution. We show that this very restrictive assumption can be weakened. Indeed, the results of Benveniste et al. for stochastic approximation with Markovian perturbations are used to establish the convergence of SAEM when it is coupled with a Markov chain Monte-Carlo procedure. This result is very useful for many practical applications. Applications to the convolution model and the change-points model are presented to illustrate the proposed method.
Mathematics Subject Classification: 62F10 / 62L20 / 65C40
Key words: EM algorithm / SAEM algorithm / stochastic approximation / MCMC algorithm / convolution model / change-points model.
© EDP Sciences, SMAI, 2004
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.