Outcome Based Matching

TL;DR

This paper introduces a novel metric learning approach for outcome-based matching in high-dimensional data, aiming to improve the accuracy of treatment effect estimates by effectively weighting relevant covariates.

Contribution

It presents a new method for learning a covariate metric that reduces variance in treatment effect estimation by emphasizing outcome-related features.

Findings

Reduced variance in treatment effect estimates

Effective covariate weighting for matching

Improved estimation accuracy in high-dimensional data

Abstract

We propose a method to reduce variance in treatment effect estimates in the setting of high-dimensional data. In particular, we introduce an approach for learning a metric to be used in matching treatment and control groups. The metric reduces variance in treatment effect estimates by weighting covariates related to the outcome and filtering out unrelated covariates.