\magnification=1200
\baselineskip=20pt
\nopagenumbers
\font\big=cmr12 scaled \magstep2

\centerline{\bf STANFORD UNIVERSITY}
\centerline{\bf DEPARTMENT OF STATISTICS}
\centerline{\big DEPARTMENT SEMINAR}
\bigskip

\baselineskip=12pt
\centerline{4:15 p.m., Tuesday, August 27, 2002}
\centerline{Sequoia Hall Room 200}
\centerline{(Cookies at 3:45 in 1st Floor Lounge)}

\bigskip
\baselineskip=15pt
\centerline{\sl Eugenio Regazzini}
\centerline{\sl Universita' degli Studi di Pavia, Italia}
\bigskip
\centerline{\bf Suggestions for the reconciliation of theories in statistics}
\bigskip

An atypical approach to statistical inference is discussed, which: a)
preserves 
the essential peculiarities of the Bayesian dynamics of the process of
learning 
from experience; (b) does not require the assessment of any prior
distribution 
for unobservable parameters; (c) is compatible with any statistical
model. Since 
the methods, that follow from such an approach, are implemented by
optimizing 
the substitution of a predictive distribution consistent with a specific
model, 
by that very same  model, they will be called predictive self
controlling. These 
methods are applied to the estimation of the parameters of a few
remarkable 
models, on the basis of real data sets. A careful study of consistency
and 
robustness of these estimates are presented. Some preliminary results on
their 
asymptotic distribution are also given.


\bye