|
Activity Number:
|
101
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #307035 |
|
Title:
|
Bayesian Analysis of Longitudinal Binary Data Using Markov Regression Models with Skewed Links
|
|
Author(s):
|
Seongho Song and Younshik Chung*+ and Dipak Dey and Alaattin Erkanli
|
|
Companies:
|
University of Cincinnati and Pusan National University and University of Connecticut and Duke University Medical Center
|
|
Address:
|
30 JangJeon-dong, Geumjeong-gu, Pusan, 609-735, Republic of Korea
|
|
Keywords:
|
Bayesian Markov regression model ; correlated Bernoulli process ; skewed link ; latent variables ; reversible jumps MCMC
|
|
Abstract:
|
In this research, we consider nonhomogeneous Markov regression models of unknown order as a means to assess the duration of autoregressive dependence in longitudinal binary data. We describe a subject's transition probability evolving over time using logistic regression models for his/her past outcomes and covariates. Our main goal is to develop the appropriate probability model for the correlated Bernoulli process in the presence of covariate information. In this model, we consider the skewed links for the link function given by Chen, Dey, and Shao (1999). Then, the model parameters order of transitions are estimated using reversible jump Markov chain Monte Carlo (MCMC) approach (Green 1995; Green and Richardson 1997).
|
