pblm - Bivariate Additive Marginal Logistic Regression via Maximum Penalized Likelihood Estimation

Bivariate additive categorical regression for moderate-to-small size datasets. Under a multinomial scheme, it is possible to fit bivariate models when the two responses are nominal, ordinal or mixed nominal/ordinal. Partial proportional odds models with (non-)uniform association structure can be fitted with the possibility to specify several logit types and parametrizations for the marginals and the association, including the Dale's model. The association structure can also be smoothed using penalty terms of polynomial type. P-splines are used in the additive part of the model. Common methods such as summary, residuals and predict are available.