All
Contents Copyright.Last revised: 20/06/2014
Ioannis Ntzoufras
Publications Page
(previous title: Marginal Likelihood Estimation from a Single Run Metropolis-Hastings Output
for Models with Local Independence)
Silia Vitoratou, Ioannis Ntzoufras and Irini Moustaki (2014)
The marginal likelihood, can be notoriously difficult to compute, and particularly so in high dimensional problems. Chib and Jeliazkov employed the local reversibility of the Metropolis-Hastings algorithm to construct an estimator in models where full conditional densities are not available analytically. The estimator is free of distributional assumptions and is directly linked to the simulation algorithm. However, it generally requires a sequence of reduced Markov chain Monte Carlo (MCMC) runs which makes the method computationally demanding especially in cases when the parameter space is large. In this article, we study the implementation of this estimator on latent variable models which embed independence of the responses to the observables given the latent variables (conditional or local independence). This property is employed in the construction of a multi-block Metropolis-within-Gibbs algorithm that allows to compute the estimator in a single run, regardless of the dimensionality of the parameter space. The counterpart one-block algorithm is also considered here, by pointing out the difference between the two approaches. The paper closes with the illustration of the estimator in simulated and real life data sets.
Keywords: Generalised linear latent variable models, mixed effects models, bridge sampling, Monte Carlo error, Laplace-Metropolis estimator.Download:
Article in JSCS 11/11/2013 can be found here
Final Version: 14/1/2013 available here.
First version: 24/10/2011 available here.
«Η παρούσα έρευνα χρηματοδοτήθηκε από τους πόρους του Ειδικού Λογαριασμού Κονδυλίων Έρευνας του Οικονομικού Πανεπιστημίου Αθηνών».
«This research was funded by the Research Centre of the Athens University of Economics and Business».
All
Contents Copyright.