Joint Bayesian Analysis of Multiple Response-Types Using the Hierarchical Generalized Transformation Model

Consider the situation where an analyst has a Bayesian statistical model that performs well for continuous data. However, suppose observed dataset consists of multiple response-types (e.g., continuous, count-valued, Bernoulli trials, etc.), which are distributed from more than one class distributions. We refer to these types data as “multiple response-type” datasets. The goal this article is introduce reasonable easy-to-implement all-purpose method “converts” responses (call preferred model) into response-type To do this, we consider transformation data, such transformed can be reasonably modeled using model. What unique with our strategy treat transformations unknown and use approach uncertainty. implementation straightforward, involves two steps. first step produces posterior replicates latent conjugate multivariate (LCM) second generating values distribution implied by demonstrate flexibility through application additive regression trees (BART) spatio-temporal mixed effects (SME) provide thorough joint analysis coronavirus disease 2019 (COVID-19) cases, adjusted closing price Dow Jones Industrial (DJI), Google Trends