Presentation Title

Bayesian Survival Analysis for Mean Residual Life Functions under the Exponeniated Weibull Regression Model

Start Date

November 2016

End Date

November 2016

Location

HUB 302-#183

Type of Presentation

Poster

Abstract

The mean residual life function provides the probability of the expected remaining life of a subject given survival up to a certain time. The mean residual life function is of interest for a variety of applications in reliability, econometrics, biostatistics, etc. The mean residual life function uniquely characterizes the survival distribution, however, traditional survival models tend to be restrictive in its shape for the corresponding mean residual life function. Hence for this research study, we focus on the Exponentiated Weibull model, whose mean residual life function can take on a variety of shapes. We apply Bayesian methods to obtain inference on a data set describing the survival time of two experimental groups of rats: ad libitum (free-eating) group and restricted group. In order to extract maximal information from the dataset, we develop a regression extension to the Exponentiated Weibull model through the scale parameter. This allows us to borrow strength across experimental groups increasing the posterior precision. We demonstrate this property by comparing the results under the regression model to the results obtained via fitting the groups independently.

This document is currently not available here.

Share

COinS
 
Nov 12th, 4:00 PM Nov 12th, 5:00 PM

Bayesian Survival Analysis for Mean Residual Life Functions under the Exponeniated Weibull Regression Model

HUB 302-#183

The mean residual life function provides the probability of the expected remaining life of a subject given survival up to a certain time. The mean residual life function is of interest for a variety of applications in reliability, econometrics, biostatistics, etc. The mean residual life function uniquely characterizes the survival distribution, however, traditional survival models tend to be restrictive in its shape for the corresponding mean residual life function. Hence for this research study, we focus on the Exponentiated Weibull model, whose mean residual life function can take on a variety of shapes. We apply Bayesian methods to obtain inference on a data set describing the survival time of two experimental groups of rats: ad libitum (free-eating) group and restricted group. In order to extract maximal information from the dataset, we develop a regression extension to the Exponentiated Weibull model through the scale parameter. This allows us to borrow strength across experimental groups increasing the posterior precision. We demonstrate this property by comparing the results under the regression model to the results obtained via fitting the groups independently.