[Forum SIS] Avviso di seminario - Marco Minozzo, Universita' degli Studi di Verona

[Maria Giovanna Ranalli]

.: Prof. Marco Minozzo, Dipartimento di Economie, Societa' ed  
Istituzioni, Sezione di Statistica, Universita' degli Studi di Verona :.

Venerdi 4 Dicembre, ore 12
Dipartimento di Economia, Finanza e Statistica, Aula 101
Università degli Studi di Perugia
.: Some Advances in Multivariate Model-Based Geostatistics :.

.: Abstract :.
In geostatistics the analysis of multivariate spatial data is
traditionally carried out resorting to the linear model of
coregionalization (Wackernagel, 1995; Chiles and Delfiner, 1999)
and to the "factorial kriging analysis" (Matheron, 1982).

Although these analyses, and the implied predictions based on
factorial cokriging, neglect any consideration about the form
of the statistical distribution of the data, and are based solely
on the spatial dependence in the data measured by variograms and
cross-variograms, they are perfectly adequate, in the sense of
providing predictions that are unbiased and with minimum mean
square error, in the case in which the data can be assumed to
follow a (multivariate) Gaussian distribution.

On the other hand, in the case of count data, of asymmetric data,
or, in general, of non-Gaussian data, as, for instance, in many
environmental, epidemiological, or business applications, these
analyses can lead to highly inefficient predictions.

On the lines of the most recent proposals appeared in the
literature for the modelling of univariate non-Gaussian spatial
data (Diggle, Moyeed and Tawn, 1998; Diggle and Ribeiro Jr., 2007),
we will present some advances in the modelling of multivariate
geostatistical data using a hierarchical spatial factor model built
on a variant of the classical linear model of coregionalization.
In particular, we will report some advances in the estimation of
the parameters of this model through stochastic EM algorithms
(Fort and Moulines, 2003) and simulated likelihoods (Geyer, 1994;
Christensen, 2004).
Cordiali saluti

M. Giovanna Ranalli

~ Dipartimento di Economia, Finanza e Statistica
~ Sezione di Statistica
~ Via Pascoli
~ 06123 Perugia - Italy

~ Tel +39 075 5855939
~ Fax +39 075 5855950
~ url: www.stat.unipg.it/~giovanna

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.stat.unipg.it/pipermail/sis/attachments/20091128/d8207d77/attachment.html>