A0342
Title: A Bayesian approach for inference on mixed graphical models
Authors: Marina Vannucci - Rice University (United States) [presenting]
Abstract: Mixed data refers to a type of data in which variables can be of multiple types, such as continuous, discrete, or categorical. This data is routinely collected in various fields, including healthcare and social sciences. A common goal in the analysis of such data is to identify dependence relationships between variables, for an understanding of their associations. A Bayesian pairwise graphical model is proposed that estimates conditional independencies between any type of data. A flexible modeling construction is implemented, which includes zero-inflated count data and can also handle missing data. It is shown that the model maintains both global and local Markov properties. A spike-and-slab prior is employed for the estimation of the graph, and an MCMC algorithm is implemented for posterior inference based on conditional likelihoods. Performances on simulated data are assessed, and results are compared with existing methods. Finally, an analysis of real data from adolescents diagnosed with an eating disorder is presented.
