|
Activity Number:
|
223
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, August 1, 2016 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
|
Abstract #317987
|
View Presentation
|
|
Title:
|
Bayesian Nonparametric Modeling and Inference for Mean Residual Life Functions
|
|
Author(s):
|
Athanasios Kottas* and Valerie Poynor
|
|
Companies:
|
University of California at Santa Cruz and California State University at Fullerton
|
|
Keywords:
|
Bayesian nonparametrics ;
Dependent Dirichlet process ;
Dirichlet process mixtures ;
Markov chain Monte Carlo ;
Survival regression
|
|
Abstract:
|
The mean residual life function provides the expected remaining lifetime given survival up to a particular time. Owing to its ready interpretation and the fact that it characterizes the survival distribution, this functional is of direct interest in biomedical applications. We will present an approach to Bayesian inference for mean residual life functions built from a nonparametric mixture model for the associated survival distribution. We address the importance of careful kernel selection to ensure desirable properties for the mean residual life function arising from the mixture distribution. The practically relevant extension to inference for mean residual life functions in the presence of covariates will also be addressed. A feature of the proposed nonparametric regression models is that they achieve flexible shapes for the conditional survival distribution given the covariates while retaining structure in the implied prior model for the conditional mean residual life function. The methods will be illustrated with simulated and real data examples.
|
Authors who are presenting talks have a * after their name.
