compboost: Modular Framework for Component-Wise Boosting
Daniel Schalk, Janek Thomas, Bernd Bischl · The Journal of Open Source Software · 2018
In high-dimensional prediction problems, especially in the p ≫ n situation, feature selection is an essential tool.A fundamental method for problems of this type is componentwise gradient boosting, which automatically selects from a pool of base learners -e.g.simple linear effects or component-wise smoothing splines (Schmid & Hothorn, 2008) -and produces a sparse additive statistical model.Boosting these kinds of models maintains interpretability and enables unbiased model selection in high-dimensional feature spaces (Hofner, Hothorn, Kneib, & Schmid, 2012).