We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ME

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Methodology

Title: Kernel Methods for Multistage Causal Inference: Mediation Analysis and Dynamic Treatment Effects

Abstract: We propose simple estimators for mediation analysis and dynamic treatment effects over short horizons based on kernel ridge regression. We study both nonparametric response curves and semiparametric treatment effects, allowing treatments, mediators, and covariates to be continuous or discrete in general spaces. Our key innovation is a new RKHS technique called sequential mean embedding, which facilitates the construction of simple estimators for complex causal estimands, including new estimands without existing alternatives. In particular, we propose machine learning estimators of dynamic dose response curves and dynamic counterfactual distributions without restrictive linearity, Markov, or no-effect-modification assumptions. Our simple estimators preserve the generality of classic identification while also achieving nonasymptotic uniform rates for causal functions and semiparametric efficiency for causal scalars. In nonlinear simulations with many covariates, we demonstrate state-of-the-art performance. We estimate mediated and dynamic response curves of the US Job Corps program for disadvantaged youth, and share a data set that may serve as a benchmark in future work.
Comments: 88 pages. Material in this draft previously appeared in a working paper presented at the 2020 NeurIPS Workshop on ML for Economic Policy (arXiv:2010.04855v1). We have divided the original working paper (arXiv:2010.04855v1) into two projects: one paper focusing on static settings (arXiv:2010.04855) and this paper focusing on dynamic settings
Subjects: Methodology (stat.ME); Machine Learning (cs.LG); Econometrics (econ.EM); Machine Learning (stat.ML)
Cite as: arXiv:2111.03950 [stat.ME]
  (or arXiv:2111.03950v3 [stat.ME] for this version)

Submission history

From: Rahul Singh [view email]
[v1] Sat, 6 Nov 2021 19:51:39 GMT (166kb,D)
[v2] Tue, 23 Aug 2022 15:09:16 GMT (185kb,D)
[v3] Mon, 5 Dec 2022 00:19:05 GMT (210kb,D)

Link back to: arXiv, form interface, contact.