Boosting with copula-based components

Abstract The authors propose new additive models for binary outcomes, where the com-ponents are copula-based regression models (Noh et al., 2013), and designed suchthat the model may capture potentially complex interaction effects. The modelsdo not require discretisation of continuous covariates, and are therefore suit-able for problems with many such covariates. A fitting algorithm, and efficientprocedures for model selection and evaluation of the components are described.Software is provided in the R-package copulaboost. Simulations and illustrationson data sets indicate that the method’s predictive performance is either betterthan or comparable to the other methods.