|
Activity Number:
|
76
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 8:00 PM to 9:50 PM
|
|
Sponsor:
|
General Methodology
|
| Abstract - #305686 |
|
Title:
|
A Bayesian Approach to Semicontinuous Longitudinal Data
|
|
Author(s):
|
Bing Han*+ and Wei Huang
|
|
Companies:
|
The Pennsylvania State University and Temple University
|
|
Address:
|
333 Thomas Building, University Park, PA, 16802,
|
|
Keywords:
|
semicontinuous data ; panel data ; Bayes hierarchical model ; two-parts model
|
|
Abstract:
|
Semicontinuous data describes the mixture in response of a continuous component and a degenerate component. The classical models on this type of data consist of two-parts model (Olsen and Schafer 2001) and Heckman selection model (Heckman 1977). We adapted the classical two-parts model to a hierarchical Bayesian structure, by which longitudinal correlation is modeled by random effects. An MCMC procedure is developed and several optimizations for MCMC, including blocking parameter space and missing indicators, are discussed. We further incorporated a Bayes mixture as the top hierarchical level. The mixture identifies subjects in panel data into different classes corresponding to different responsive behavior. Finally, the methodology has been applied to the case study of R&D cost and patent in the telecommunication industry.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2006 program |
