Boost-R: Gradient boosted trees for recurrence data

Recurrence data arise from multi-disciplinary domains spanning reliability, cyber security, healthcare, online retailing, etc. This paper investigates an additive-tree-based approach, known as Boost-R (Boosting for Data), recurrent event with both static and dynamic features. constructs ensemble of gradient boosted additive trees to estimate the cumulative intensity function process, where a new tree is added by minimizing regularized L2 distance between observed predicted intensity. Unlike conventional regression trees, time-dependent constructed on each leaf. The sum these functions, multiple yields estimator divide-and-conquer nature tree-based methods appealing when hidden sub-populations exist within heterogeneous population. non-parametric helps avoid parametric assumptions complex interactions processes Critical insights advantages are investigated through comprehensive numerical examples. Datasets computer code made available GitHub. To our best knowledge, first approach modeling large-scale feature information.