|
Activity Number:
|
132
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #303589 |
|
Title:
|
Bayesian Estimation in Nonstandard Finite Mixture Models with Application to an Exposure Data Set
|
|
Author(s):
|
Miranda L. Lynch*+ and Sally W. Thurston
|
|
Companies:
|
University of Rochester School of Medicine and Dentistry and University of Rochester
|
|
Address:
|
Dept of Biostatistics and Computational Biology, Rochester, NY, 14642,
|
|
Keywords:
|
Bayesian estimation ; Finite mixture model ; Nonstandard mixture model ; Limit of detection
|
|
Abstract:
|
Finite mixture models (FMM) represent an important class of modeling tools with applicability in a variety of data settings. Bayesian estimation utilizing MCMC methods for these models provides an important means for extending their applicability to complex modeling situations, but it also presents unique computational challenges. This work discusses application of Bayesian estimation procedures for nonstandard mixtures composed of a degenerate point mass and a nondegenerate continuous distribution. Such mixtures have found applicability in modeling zero inflated data and measurement data in which the measurements are subject to a limit of detection (LOD). Methods discussed will be demonstrated via application to simulated data and to an exposure data set in which LOD issues are present.
|
