[Forum SIS] Seminari HOSNI e REGAZZINI

[Matteo Ruggiero]

********************************************************
		        CICLO DI SEMINARI
COLLEGIO CARLO ALBERTO - MONCALIERI (TO)
********************************************************

Venerdi' 6 Maggio 2011, 
nell'ambito del ciclo di seminari organizzato dalla "de Castro Statistics Initiative" (www.carloalberto.org/stats), 
presso la Sala Rossa del Collegio Carlo Alberto, via Real Collegio 30, Moncalieri, Torino,
si terranno i seguenti seminari:


ore 11.00, HYKEL HOSNI  (Scuola Normale Superiore)

"SUBJECTIVE PROBABILITY AS A NORMATIVE THEORY OF RATIONAL BELIEF"


ore 12.00, EUGENIO REGAZZINI  (Universita' di Pavia)

"FINITELY ADDITIVE PROBABILITIES IN STATISTICS".



Seguono gli abstracts dei seminari.
Tutti gli interessati sono invitati a partecipare.



HYKEL HOSNI  (Scuola Normale Superiore)

Subjective probability as a normative theory of rational belief

Bruno de Finetti developed his subjectivistic views in an attempt to clarify -- by means of "operational definitions" -- the very meaning of  probability.  His approach, which was ultimately inspired by decision theoretic considerations, rapidly extended outside the foundations of probability and contributed to the development of the so-called bayesian view of rationality.  Today, bayesianism polarises decision theorists like possibly no other foundational issue,  opening up to much heated debates and discussions. While some  point to real difficulties in taking strict forms of subjectivism as the normative standard for rational belief, others end up being nothing more than religious disputes.
The purpose of this talk is to offer a faith-free assessment of de Finetti's key foundational views with the two-fold aim of recalling its deep strengths while revealing some of its intrinsic limitations. In particular I will suggest that under a careful characterization of bayesianims, some of the  most intensely debated criticisms loose much of their appeal. As a particularly important case in point, I will refer to the much discussed inadequacy of bayesianism to capture rational reasoning under ignorance (otherwise referred to as reasoninig with ambiguity).


EUGENIO REGAZZINI  (Universita' di Pavia)

Finitely additive probabilities in statistics

It is well-known that according to de Finetti's viewpoint the set of admissible probability laws also includes those which are finitely additive, but not countably additive. Does such an extension yield relevant consequences for statistics? An attempt to answer this question is made by resorting to examples, which concern, on the one hand, usual limiting properties of frequentist statistical procedures and, on the other hand, features of the trajectories of random probabilities commonly used in Bayesian statistics.


---------
Matteo Ruggiero
University of Turin
http://web.econ.unito.it/ruggiero





-------------- parte successiva --------------
Un allegato HTML è stato rimosso...
URL: <http://www.stat.unipg.it/pipermail/sis/attachments/20110502/94bedaab/attachment.html>