Debamita Kundu
University of Virginia
This presentation describes the development of novel Bayesian methodologies tailored for multivariate issues within public health data, employing shrinkage priors.
Covariance estimation for multiple groups is pivotal for drawing inference from a heterogeneous population. One should seek to share information about common features in the dependence structures across the various groups. Our contribution lies in introducing a novel approach for estimating the covariance matrices for multiple groups using a Bayesian hierarchical sparse factor (BaSH-F) model. We apply our methodology to the NICHD Consecutive Pregnancies study to estimate the correlations between birth weights and gestational ages of three consecutive births within four different subgroups (underweight, normal, overweight and obese) of women.
In another application, we have used a “global-local” shrinkage prior in a way that shares the information between the main effects and interaction effects for estimating the complex relationship between environmental chemical exposures and disease risk. In the presence of a large number of chemicals, it is difficult to estimate the interactive effects without incorporating reasonable prior information. In this approach, we incorporate the dependence structure between main effects and interaction effects through a global-local shrinkage prior framework. We apply this methodology to the NCI-SEER NHL study, investigating associations between pesticide exposures and the risk of non-Hodgkin lymphoma (NHL).
