Easily Computed Marginal Likelihoods from Posterior Simulation Using the THAMES Estimator

Orateur: Marie Perrot-Dockès
Localisation: ,
Type: Séminaire de probabilités et statistiques
Site: UGE , 4B 125
Date de début: 21/12/2023 - 10:30
Date de fin: 21/12/2023 - 12:00

We propose an easily computed estimator of marginal likelihoods from posterior simulation output, via reciprocal importance sampling, combining earlier proposals of DiCiccio et al (1997) and Robert and Wraith (2009). This involves only the unnormalized posterior densities from the  sampled parameter values, and does not involve additional simulations beyond the main posterior simulation, or additional complicated calculations, provided that the parameter space is unconstrained. Even in this case, it is easily adjusted by a simple Monte Carlo approximation. It is unbiased for the reciprocal of the marginal likelihood, consistent, has finite variance, and is asymptotically normal. It involves one user-specified control parameter, and we derive an optimal way of specifying this. We illustrate it with several numerical examples.