0

BoXHED2.0: Scalable boosting of dynamic survival analysis

Modern applications of survival analysis increasingly involve time-dependent covariates. The Python package BoXHED2.0 is a tree-boosted hazard estimator that is fully nonparametric, and is applicable to survival settings far more general than right-censoring, including recurring…

Open
Year
2021
ArXiv
arxiv.org/abs/2103.12591
Hosting
Full text hostedCC-BY-4.0

Cite

Notes

Only stored in your browser.

Attribution

Abstract & full text
arxiv.org/abs/2103.12591CC-BY-4.0
TL;DR
Semantic Scholar
Attribution policy →

Abstract

Modern applications of survival analysis increasingly involve time-dependent covariates. The Python package BoXHED2.0 is a tree-boosted hazard estimator that is fully nonparametric, and is applicable to survival settings far more general than right-censoring, including recurring events and competing risks. BoXHED2.0 is also scalable to the point of being on the same order of speed as parametric boosted survival models, in part because its core is written in C++ and it also supports the use of GPUs and multicore CPUs. BoXHED2.0 is available from PyPI and also from www.github.com/BoXHED.