THE GEOFFREY S. WATSON LECTURE SERIES
Geoffrey S. Watson was previously chair
of the Department of Statistics in the School of Arts and Sciences at
Johns Hopkins University. In that role, he stimulated active interactions
among statisticians and biostatisticians across the Hopkins campuses. In
honor of his contributions to statistical sciences, the Departments of
Mathematical Sciences in the School of Arts and Sciences and Biostatistics
in the School of Public Health have created an annual Watson Lecture to be
jointly sponsored in the spring term of each academic year. A joint
committee of faculty will select the lecturer, who will present a research
lecture that honors the spirit of original investigation typified by Geoff
Watson. An honor roll of past Watson Lecturers will be presented at each
year’s meeting.
1999 Watson
Lecturer:
Rudolf Beran,
UC-Berkeley Department of Statistics
April 14, 1999
"Smoothing in the Linear Model"
1998 Watson Lecturer:
Susan Murphy,
University of Michigan
"Profile Likelihood Inference in High-Dimensional Models"
March 11, 1998
ABSTRACT:
In high dimensional models, interest often lies primarily in a vector
parameter and the nuisance parameter is a function, such as a distribution
function or a smooth regression function. This talk concerns the
verification of the intuitive practice of profiling the nuisance parameter
out of the likelihood and using the resulting profile likelihood as if it
is a likelihood for the vector parameter. That is, will maximizing the
profile likelihood yield an asymptotically normal estimator of the
parameter of interest? Can the profile likelihood be used to make
likelihood ratio tests and confidence intervals? Can minus the second
derivative matrix of the profile likelihood be used to estimate the
information matrix? In the parametric setting these properties follow
from
a quadratic approximation to the likelihood. This work give sufficient
conditions for the above approximation to hold in a semiparametric setting
and shows how the quadratic approximation leads to an affirmative answer
to
the above questions. The conditions are sufficiently simple so as to be
satisfied in a variety of semiparametric models, for example, the
proportional hazards model for current status data, the proportional odds
model for right censored survival data, gamma frailty model, errors in
variables for a logistic regression model, and a semiparametric logistic
regression.
Return
to Events & News List |
Return to Home Page
|
