Center for Statistical Sciences Seminar
Department of Biostatistics, Harvard University | |
Abstract: Modern cancer treatments have substantially improved cure rates and have generated a great interest in and need for proper statistical tools to analyze survival data with non-negligible cure fractions. Data with cure fractions are often complicated by dependent censoring, and analysis of this type of data typically involves untestable assumptions about the dependence of the censoring and the survival times. Motivated by the analysis of NCI SEER data, we propose a class of general semiparametric transformation cure models that allows for dependent censoring without making parametric assumptions on the dependence relationship and use the proposed methods to investigate potential racial disparities in prostate cancer cures. The proposed class of models encompasses a number of common models for the latency survival function, including the proportional hazards and proportional odds models.
***SPECIAL*** Brown Analysis Seminar
Abstract: There is a very well developed theory of uniform rectifiability in R^n (due to David, Semmes, Jones and others). There are now several results of similar nature for metric spaces. We will recall/explain several Euclidean results and then go on to their counter-parts for a general metric space. We will also discuss some parts of the Euclidean theory we have been unable (thus far) to transfer to the category of metric spaces and explain what the obstacles seem to be.
Scientific Computing Seminar
Abstract:
We present a new data structure, adapted to front propagation problems.
This structure is used here together with a variant of the antidiffusive
"Ultra Bee" algorithm on a regular mesh, leading to a fast narrow band method.
This method is tested in 2 to 4 space dimensions, and compared with the level
set approach.We then turn on academic examples for deterministic controls
problems (corresponding to a particular class of Hamilton-Jacobi-Bellman
equations) such as computing reachable sets or minimal time functions.
We finaly give a general procedure in order to deal furthermore with state
constraints on the trajectories (without using any controlability assumptions
on the dynamics on the boundary of the admissible domain).
Joint work with H. Zidani, E. Cristiani, N. Forcadel
Center for Statistical Sciences Seminar
Abstract:
We consider assessment of nonresponse bias for the mean of a survey variable Y
subject to nonresponse. We assume that there are a set of covariates observed
for nonrespondents and respondents. To reduce dimensionality and for simplicity
we reduce the covariates to a proxy variable X that has the highest correlation
with Y, estimated from a regression analysis of respondent data. We consider
adjusted estimators of the mean of Y that are maximum likelihood for a
pattern-mixture model with different mean and covariance matrix of Y and X for
respondents and nonrespondents, assuming missingness is an arbitrary function
of a known linear combination of X and Y. We propose a taxonomy for the
evidence concerning bias based on the strength of the proxy and the deviation
of the mean of X for respondents from its overall mean, propose a sensitivity
analysis, and describe Bayesian versions of this approach. We propose using
the fraction of missing information from multiple imputation under the
pattern-mixture model as a measure of nonresponse bias. Methods are
demonstrated through simulation and with data from the third National Health
and Nutrition Examination Survey (NHANES III).
Candidate for Assistant Professor in the
Biostatistics Section of the Program in Public Health
PDE Seminar
Abstract: We will discuss two-dimensional traveling stratified water waves propagating over an impermeable flat bed and with a free surface. The wave's motion is assumed to be driven by surface tension on the upper boundary and a gravitational force acting on the body of the fluid. Such waves are commonly seen to form when, for example, a wind blows over a quiescent body of water. We shall present some new results on the existence of global continua of classical solutions of this type. In the process, we shall also answer some open questions for the constant density case.