Online Program Home
My Program

Abstract Details

Activity Number: 522 - Contributed Poster Presentations: Biometrics Section
Type: Contributed
Date/Time: Wednesday, July 31, 2019 : 10:30 AM to 12:20 PM
Sponsor: Biometrics Section
Abstract #306369
Title: Bivariate Hierarchical Bayesian Model for Combining Estimates from Multiple Sources and Domains
Author(s): Yujing Yao* and Todd Ogden and Qixuan Chen
Companies: Columbia University and Columbia University and Columbia University
Keywords: Brain Imaging; Hierarchical Bayesian model; Multivariate model; Small area estimation

In many situations, one is interested in combining estimates from multiple sources or domains given corresponding measures of the uncertainty of each estimate. Efficient inference can be accomplished based on a hierarchical Bayesian model with appropriate priors. Such a model should take into account the uncertainty of the estimates, the uncertainty in the measures of the uncertainty, and the association between them. In this paper, a general bivariate hierarchical Bayesian model for both estimates and their corresponding measures of uncertainty is proposed. This approach can be thought of as an improvement to the general univariate hierarchical Bayesian model in terms of estimation accuracy and efficiency in the situation in which the measure and its uncertainty are correlated. We carry out Monte Carlo simulation studies to compare the performance of these two hierarchical Bayesian models under different scenarios and present estimation results. We illustrate the applications of the bivariate hierarchical Bayesian model with examples in small area estimation and brain image studies using PET data, but the model is also applicable to applications in meta-analysis and other fields.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program