Build gradient boosted regression trees that fit sequential residuals and combine learners with a configurable learning rate.
