[CCoE Notice] Industrial Engineering Seminar Announcement

[Tekin, Eylem]

INDUSTRIAL ENGINEERING SEMINAR

Date: March 29, 2013
Time: 10am-11am
Place: D 102

Biased Sampling and Failure Time Data Analysis
By

Dr. Hao Liu
Associate Professor of Biostatistics
Dan L. Duncan Cancer Center and Department of Medicine
Baylor College of Medicine

Abstract

                When a recorded observation does not represent the distribution of the targeted population, biased sampling occurs and statistical methodology is required to correct the bias.    Biased sampling has wide applications from astronomical survey, clinical trials, epidemiology study, financial performance, to industrial engineering.   In this talk, we review the relevant statistical methods with the focus on the analysis of failure time data.  By modeling a general sample selection mechanism parametrically, we show that the efficient semiparametric maximum likelihood estimation (MLE) can be achieved for the popular Cox model.  We present a novel expectation and maximization (EM) algorithm to find the semiparametric MLE, which has considerable computational advantages.  Finally, we demonstrate the proposed procedure with a real data analysis in an epidemiology study.

Biography


Dr. Hao Liu is an Associate Professor in Biostatistics at the Dan L. Duncan Cancer Center and Department of Medicine at Baylor College of Medicine.  He received his undergraduate degree in Mathematics from Peking University in China, and a Master degree in Biostatistics from the University of California at Berkeley.  Dr. Liu finished his doctoral study in Biostatistics at the University of Washington at Seattle.  After a stint as an Assistant Professor at the University of California at Davis, he moved to Houston to joint Baylor College of Medicine.  Dr. Liu has research interests on clinical trials as well as general statistical methodology.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://Bug.EGR.UH.EDU/pipermail/engi-dist/attachments/20130322/3077d7d5/attachment.html