Pulling back the curtain: the road from statistical estimand to machine-learning based estimator for epidemiologists (no wizard required)

• DOI: 10.48550/arxiv.2502.05363
• Publication Date: 2025-02-07
• Authors: Audrey Renson; Lina Montoya; Dana E. Goin; Iván Dı́az; Rachael Ross
• Keywords: Wizard

Epidemiologists increasingly use causal inference methods that rely on machine learning, as these approaches can relax unnecessary model specification assumptions. While deriving and studying asymptotic properties of such estimators is a task usually associated with statisticians, it useful for epidemiologists to understand the steps involved, are often at forefront defining important new research questions translating them into parameters be estimated. In this paper, our goal was provide relatively accessible guide through process (i) an estimator based so-called efficient influence function (which we define explain), (ii) showing estimator's ability validly incorporate by demonstrating rate double robustness property. The derivations in paper mainly algebra some foundational results from statistical inference, which explained.