Sufficient dimension reduction for average causal effect estimation

DOI10.1007/S10618-022-00832-5zbMATH Open1497.62049arXiv2009.06444OpenAlexW3086776987MaRDI QIDQ2147407FDOQ2147407

J. Liu, Jiuyong Li, Thuc Duy Le, Debo Cheng, Kui Yu, Lin Liu

Publication date: 20 June 2022

Published in: Data Mining and Knowledge Discovery (Search for Journal in Brave)

Abstract: Having a large number of covariates can have a negative impact on the quality of causal effect estimation since confounding adjustment becomes unreliable when the number of covariates is large relative to the samples available. Propensity score is a common way to deal with a large covariate set, but the accuracy of propensity score estimation (normally done by logistic regression) is also challenged by large number of covariates. In this paper, we prove that a large covariate set can be reduced to a lower dimensional representation which captures the complete information for adjustment in causal effect estimation. The theoretical result enables effective data-driven algorithms for causal effect estimation. We develop an algorithm which employs a supervised kernel dimension reduction method to search for a lower dimensional representation for the original covariates, and then utilizes nearest neighbor matching in the reduced covariate space to impute the counterfactual outcomes to avoid large-sized covariate set problem. The proposed algorithm is evaluated on two semi-synthetic and three real-world datasets and the results have demonstrated the effectiveness of the algorithm.

Full work available at URL: https://arxiv.org/abs/2009.06444

Recommendations

• On estimating regression-based causal effects using sufficient dimension reduction
• A robust and efficient approach to causal inference based on sparse sufficient dimension reduction
• Data-driven algorithms for dimension reduction in causal inference
• Propensity score modelling in observational studies using dimension reduction methods
• Sufficient dimension reduction for feasible and robust estimation of average causal effect

zbMATH Keywords

causal inferencesufficient dimension reductioncausal effects estimationconfounding bias

Mathematics Subject Classification ID

Causal inference from observational studies (62D20)

Cites Work

• Causation, prediction, and search
• Generalized random forests
• Recursive partitioning for heterogeneous causal effects
• Estimation and Inference of Heterogeneous Treatment Effects using Random Forests
• Causal Inference for Statistics, Social, and Biomedical Sciences
• Theory of Reproducing Kernels
• Title not available (Why is that?)
• The central role of the propensity score in observational studies for causal effects
• A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect
• Matching methods for causal inference: a review and a look forward
• Kernel methods in machine learning
• Efficient semiparametric estimation of multi-valued treatment effects under ignorability
• Title not available (Why is that?)
• On the Role of the Propensity Score in Efficient Semiparametric Estimation of Average Treatment Effects
• Large Sample Properties of Matching Estimators for Average Treatment Effects
• Convergence rates of least squares regression estimators with heavy-tailed errors
• Title not available (Why is that?)
• Estimating high-dimensional intervention effects from observational data
• A new criterion for confounder selection
• Matching on the Estimated Propensity Score
• Covariate Balancing Propensity Score
• Using Multivariate Matched Sampling and Regression Adjustment to Control Bias in Observational Studies
• Graphics for Regressions With a Binary Response
• An alternative robust estimator of average treatment effect in causal inference
• Doubly robust nonparametric inference on the average treatment effect
• Numerical Issues in Statistical Computing for the Social Scientist
• A robust and efficient approach to causal inference based on sparse sufficient dimension reduction
• On estimating regression-based causal effects using sufficient dimension reduction
• Covariate selection for the nonparametric estimation of an average treatment effect
• Covariate selection strategies for causal inference: Classification and comparison
• A generalized back-door criterion
• Outcome‐adaptive lasso: Variable selection for causal inference
• Data-driven confounder selection via Markov and Bayesian networks
• Title not available (Why is that?)

This page was built for publication: Sufficient dimension reduction for average causal effect estimation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2147407)