special statistics seminar, 1st July 1pm RC-3084

Marginal likelihoods via power posteriors

• Speaker: Dr. Nial Friel, University of Glasgow
• Time: 1:00p.m. Friday 1st July
• Venue: Red Center Room RC-3084

Model choice plays an increasingly important role in Statistics. From a Bayesian
perspective a crucial goal is to compute the marginal likelihood of the data for
a given model. This however is typically a di±cult task since it amounts to
integrating over all model parameters. The aim of this paper is to illustrate
how this may be achieved using ideas from thermodynamic integration or path
sampling (Gelman and Meng 1998). We show how the marginal likelihood can
be computed via MCMC methods on modi¯ed posterior distributions. This
then allows Bayes factors or posterior model probabilities to be calculated. We
show that this approach requires very little tuning, and is straightforward to
implement. This new method is illustrated in a variety of challenging statistical
settings.

You are welcome to join us for lunch with the speaker at 12:15, meeting outside
the stats corridor on level 1 of the Red Centre.
Acting seminar co-ordinator: Scott Sisson
e-mail: Scott@maths.unsw.edu.au