EcoSta 2022: Start Registration
View Submission - EcoSta2022
A0335
Title: Estimating heterogeneous treatment effects with right-censored data via causal survival forests Authors:  Yifan Cui - Zhejiang University (China) [presenting]
Michael Kosorok - University of North Carolina at Chapel Hill (United States)
Erik Sverdrup - Standford University (United States)
Stefan Wager - Stanford University (United States)
Ruoqing Zhu - University of Illinois at Urbana-Champaign (United States)
Abstract: Forest-based methods have recently gained popularity for non-parametric treatment effect estimation. Building on this line of work, we introduce causal survival forests, which can be used to estimate heterogeneous treatment effects in survival and observational settings where outcomes may be right-censored. The approach relies on orthogonal estimating equations to robustly adjust for both censoring and selection effects. In the experiments, we find our approach to perform well relative to a number of baselines.