A profile boosting framework for feature selection in parametric models. It offers a unified interface pboost() and several wrapped models, including linear model, generalized linear models, quantile regression, Cox proportional hazards model, beta regression, spatial auto-regressive models.