|
Activity Number:
|
247
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Council of Chapters
|
| Abstract - #300212 |
|
Title:
|
Bayesian Semiparametric Modeling and Inference for Longitudinal Diagnostic Testing Data
|
|
Author(s):
|
Wesley O. Johnson*+ and Michelle Norris
|
|
Companies:
|
University of California, Irvine and University of California, Davis
|
|
Address:
|
2322 Bren Hall, Irvine, CA, 92697-1250,
|
|
Keywords:
|
Bayesian Nonparametric ; Diagnostic testing ; Longitudinal ; Dirichlet Process Mixture ; Change point
|
|
Abstract:
|
We propose a model for joint longitudinal data where one response variable is continuous and the other is categorical; both are diagnostic measures for infection. Individuals may or may not become infected during the course of the study. We include a latent time varying infection status variable for each individual. Both response variables are modeled to behave differently after infection than before. Classification is a primary goal. It is anticipated that there will be clusters of individuals with distinct response patterns in the continuous variate through time and so a Dirichlet Process Mixture model is employed to model random effects and/or functional random effects for each individual. Inferences are based on MCMC methods for varying-dimension parameter spaces. Methods are applied to ELISA and fecal culture screening data for Johne's Disease in cattle.
|
