statistics faculty at
unc-ch
_____________
301 New West
919-962-2378
Fax: 919 962-1279
-
Gordon Simons
-
Professor
Education
-
B.A. (1960), M.A. (1964), Ph.D. (1966), University of Minnesota.
-
Stanford University (1966-68);
UNC-Chapel Hill (1968- ).
Honors and Awards
Research Interests
A project of mostly theoretical interest, has been an effort to unify
some old work of J. L. Doob's, concerned with the preservation of iid
random variables under a sequence of stopping times, and some recent
work by Ignatov, of a very different sort, concerned with the
distributions of record values. A current project is the writing of a
monograph with Sándor Csörgö on the "St. Petersburg
Paradox," a
fascinating problem that engaged the interest of almost every leading
thinker of mathematical note in the eighteenth century, and which, according to
economist and Nobel laureate Paul Samuelson, "enjoys an honored corner in the
memory bank of the cultured and analytic mind." Beyond its concern with the
historical side of the subject, the monograph will attempt to resolve the
paradox through the inclusion of recent unpublished probabilistic research,
some of which is still in progress.
On a problem of ammunition rationing (with L. Shepp and Y-C Yao),
Advances in Applied Probabability, 23 (1991), 624-641.
On Steinhaus' resolution of the St. Petersburg paradox (with S.
Csörgö),
Probability and Mathematical Statistics, 14, (1993), 157-172.
Approximating
the inverse of a
symmetric positive definite matrix (with
Y-C Yao), Linear Algebra and Its Applications, 281, (1998), 97-103.
abstract.
Asymptotics when the number of
parameters tends to infinity in the Bradley-Terry model for paired
comparisons, 1999 (to appear in Annals of Statistics).
Member, ISI.
Simons' interests are in statistical inference and applied probability.
Recently, he and some graduate students worked on a sequential statistical
model for conducting clinical trials with the expressed purpose of
increasing the number of patients who receive the best treatment during
and after the testing phase of the trials. Another area of interest is a
collection of inferential questions asssociated with "binary symmetric
channels," a topic in a semiparametric context with relevance to coding
and imaging. These projects include a challenging mixture of mathematics
and numerical computations.
Estimating distortion in a binary symmetric channel consistently,
IEEE Transactions on Information Theory, 37 (1991), 1466-1470.
To the Stat Home
Page