[Forum SIS] Avviso Seminari

[Silvia Cagnone]

          ****************************************************************
              Dipartimento di Scienze Statistiche "Paolo Fortunati"
                         Via delle Belle Arti 41, 40126 Bologna

********************************************************************


                         Giovedì 27 marzo ore 14:00, Aula III

Professor*  Paolo Vidoni*
Dipartimento di Scienze Economiche e Statistiche, Università di Udine

*"Predictive densities based on composite likelihood methods"*

Whenever the computation of the joint distribution of the data is not
possible or convenient,
the classical predictive procedures are not useful since they rely on the
conditional distribution
of the future random variable given the observations, which is also not
available. In this paper,
the aim is to consider a suitable notion of composite likelihood as a
starting point for specifying
composite predictive distributions, as useful surrogates for the true
unknown predictive distribution.
Among various notions of composite likelihood, we focus on the pairwise
likelihood,
obtained as a weighted product of likelihood factors related to bivariate
marginal or conditional events. The
specification of the weights, and more generally the evaluation of the
frequentist properties of
alternative pairwise predictive distributions, is performed by considering
the mean square prediction
 error of the associated predictors and the expected Kullback-Liebler loss
of the related predictive  densities.
Finally, simple examples concerning autoregressive models are presented.



                           Giovedì 27 marzo ore 15:00, Aula III


Professor*  Ruggero Bellio*
Dipartimento di Scienze Economiche e Statistiche, Università di Udine

"*Improved random effects prediction in GLMMs*"

This talk concerns prediction of random effects, and in particular of
expected responses,
in generalized linear mixed models. The emphasis is on the construction of
prediction intervals
having conditional coverage probability close to the target nominal value.
Some theoretical
results are presented and easy-to-use formulas are applied to obtain
improved random effects
prediction intervals for commonly used models. Some examples illustrate the
methodology.
This is a joint work with Paolo Vidoni, University of Udine.


-- 
______________________________________________________

Silvia Cagnone
Dipartimento di Scienze Statistiche "Paolo Fortunati"
Universita' di Bologna
Via delle belle arti 41 - 40126  Bologna,  ITALY
Tel: +39 051 2098250  Fax: +39 051 232153
http://www2.stat.unibo.it/cagnone/
______________________________________________________
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.stat.unipg.it/pipermail/sis/attachments/20140306/d9321cf5/attachment.html>