Youngjoo Cho
Konkuk University & University of Pittsburgh
Room: M3 3127
Estimation of heterogeneous effect using ensembles for competing risks data
Estimation of heterogeneous treatment effects for uncensored data has been studied extensively. However, efforts to develop methods for estimating heterogeneous treatment effects using machine learning have begun comparatively recently for censored data, especially in the context of competing risks. In this paper, we propose a novel approach to estimating heterogeneous treatment effects with respect to cumulative incidence curves in competing risks data using random forests. The proposed methods employ doubly robust orthogonal estimating equations to adjust both censoring and treatment effects based on semiparametric efficiency theory. We illustrate our method by using data from the Radiation Therapy Oncology Group (trial 9410).