Avviso di Seminario - Universita' di Pavia

["Guido Consonni" <guido.consonni@unipv.it>]

UNIVERSITA' DEGLI STUDI DI PAVIA
DIPARTIMENTO DI ECONOMIA POLITICA E METODI QUANTITATIVI
DIPARTIMENTO DI MATEMATICA "F.CASORATI"
DOTTORATO DI RICERCA IN MATEMATICA E STATISTICA
***********************************************************************

                         AVVISO DI SEMINARIO


Giovedi' 5 febbraio, alle ore 12.00, presso il Dipartimento di Economia
Politica e Metodi Quantitativi (via San Felice 5, Pavia)

                             JEAN-MICHEL MARIN
                         (Universite' Paris IX Dauphine)

terra' un seminario dal titolo

         BAYESIAN MODELLING AND INFERENCE ON MIXTURES OF DISTRIBUTIONS

Abstract:

Today's data analysts and modelers are in the luxurious position of being
able to more closely describe, estimate, predict and infer about complex
systems of interest, thanks to ever more powerful computational methods but
also wider ranges of modelling distributions. Mixture models constitute a
fascinating illustration of these aspects: while within a parametric
family, they offer malleable approximations in non-parametric settings;
although based on standard distributions, they pose highly complex
computational challenges; and they are both easy to constrain to meet
identifiability requirements and fall within the class of ill-posed
problems. They also provide an endless benchmark for assessing new
techniques, from the EM algorithm to reversible jump methodology. In
particular, they exemplify the formidable opening provided by new
computational technologies like Markov chain Monte Carlo (MCMC) algorithms.
It is no coincidence that the Gibbs sampling algorithm for the estimation
of mixtures was proposed before (Tanner and Wong, 1987) and immediately
after (Diebolt and Robert, 1990) the seminal paper of Gelfand and Smith
(1990): before MCMC was introduced, there simply was no satisfactory
approach to the computation of Bayes estimators for mixtures of
distributions.
Bayesian approaches to mixture modelling have attracted great interest
among researchers and practitioners alike. The goal of my talk is to
introduce the reader to the construction, prior modelling, estimation and
evaluation of finite mixture distributions in a Bayesian paradigm. And then
show that mixture distributions provide a flexible, parametric framework
for statistical modelling and analysis.




-- 
Guido Consonni
University of Pavia
Dipartimento Econ Pol Mat Quant.
Via S. Felice, 7
27100 Pavia (ITALY)

+39 0382 50 6225 (office)
+39 0382 30 42 26 (fax)

guido.consonni@unipv.it

http://economia.unipv.it/gconsonni