Bayesian discrete conditional transformation models (BDCTMs) provide an overarching model framework for situations where e.g. count hurdle or (non-)proportional odds models with nonlinear (interaction) effects is due. Inference via MCMC is based on the No-U-Turn sampler.
Nonlinear transformation model of patent citation counts with possibly nonlinear hurdle effects at zero.
- nonlinear conditional count transformation model with nonlinear hurdle effects
Nonlinear partial proportional odds model on forest defoliation categories with random and spatial effect
- spatial tensor spline
- nonlinear non-proportional odds
