Bayesian discrete conditional transformation models

We propose a novel Bayesian model framework for discrete ordinal and count data based on conditional transformations of the responses. The transformation function is estimated from in conjunction with an priori chosen reference distribution. For responses, resulting sense that it fully parametric yet distribution-free approach can additionally account excess zeros additive specifications. categoric our cumulative link allows inclusion linear non-linear covariate effects be made category-specific, (non-)proportional odds or hazards models more, depending choice Inference conducted by generic modular Markov chain Monte Carlo algorithm where multivariate Gaussian priors enforce specific properties such as smoothness functional effects. To illustrate versatility models, applications to counts patent citations presence treating forest health categories partial proportional are presented.