|
Activity Number:
|
286
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
ENAR
|
| Abstract - #309855 |
|
Title:
|
Bayesian Inferences for Bivariate Positive Stable Frailty
|
|
Author(s):
|
Madhuja Mallick*+ and Nalini Ravishanker and Nandini Kannan
|
|
Companies:
|
Merck & Co., Inc. and University of Connecticut and University of Texas at San Antonio
|
|
Address:
|
126 E Lincoln Avenue, Rahway, NJ, 07065,
|
|
Keywords:
|
dependence ; baseline hazard
|
|
Abstract:
|
The univariate shared frailty models have several drawbacks. In many applications, the situation arises where the subjects in the same group possess different frailty random components rather than sharing similar frailty components within the group. The notion of multivariate frailty enables us to overcome the difficulties of univariate shared frailty models. This presentation proposes bivariate positive stable frailty model to incorporate more heterogeneity than the shared positive stable frailty model. The estimation procedure for bivariate positive stable density is not easy due to lack of closed form of the bivariate positive stable density function. The Bayesian inference for the bivariate positive stable frailty model for multivariate survival data is described here.
|
