|
Activity Number:
|
252
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 13, 2002 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Stat. Sciences*
|
| Abstract - #301155 |
|
Title:
|
Bayesian Inference in Applications of Singular Value Decomposition to Biology
|
|
Author(s):
|
Katechan Jampachaisri*+ and S. Press+
|
|
Affiliation(s):
|
University of California, Riverside and University of California, Riverside
|
|
Address:
|
Department of Statistics, Riverside, California, 92521-0122, USA , Riverside, California, 92521-0122, USA
|
|
Keywords:
|
Singular value decomposition ; Bayesian univariate regression ; Bayesian classification
|
|
Abstract:
|
Many applications in biology have posed some difficulties in statistical modeling and inferences when the number of available predictors, p, is much larger than the sample size, n (p >> n). The objective of this study (part of my Ph.D. thesis) is to demonstrate the use of singular-value decomposition (SVD) in the explanatory matrices of Bayesian univariate regression and Bayesian classification with more than two categories, and to investigate its effect on posterior inferences with respect to three types of prior knowledge: vague priors, conjugate priors, and g-priors. The effect of SVD is inherent in the likelihood and is induced in the posterior density, resulting in massive dimension reduction in the regression coefficients and classification parameters. For each developed model, Markov Chain Monte Carlo, with Gibbs sampling, is implemented with the biological data to obtain the posterior distributions of interest.
|
- 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 2002 program |
