Double/Debiased/Neyman Machine Learning of Treatment Effects

摘要

Chernozhukov et al. (2016) provide a generic double/debiased machine learning (ML) approach for obtaining valid inferential statements about focal parameters, using Neyman-orthogonal scores and cross-fitting, in settings where nuisance parameters are estimated using ML methods. In this note, we illustrate the application of this method in the context of estimating average treatment effects (ATE) and average treatment effects on the treated (ATTE) using observational data. Empirical illustrations and code are available as supplementary material to this paper, and more general discussion and references to the existing literature are available in Chernozhukov et al. (2016).