标题： 构建目标最小损失/最大似然估计量：一个简单的示例以建立直觉 显示英文标题

标题： Constructing targeted minimum loss/maximum likelihood estimators: a simple illustration to build intuition

摘要： 使用机器学习来估计扰动函数（例如结果模型、倾向得分模型）在因果推断中使用的估计量中越来越常见，因为它可以减轻模型误指带来的偏差。 然而，实现有效的推断（例如，估计有效的置信区间）可能会很具挑战性。 有效影响函数（EIF）提供了一种方法，从与我们的因果问题相关的统计估计量，到一个可以有效结合机器学习的估计量。 我们的配套论文，Renson 等 2025（arXiv:2502.05363），提供了 EIF 的详细但易于理解的描述，以及从独特的统计估计量到开发一种基于 EIF 的估计量——所谓的一步估计量的步骤指南。 另一种常用的基于 EIF 的估计量是目标最大似然/最小损失估计量（TMLE）。 TMLE 的构建在统计文献中有详细的讨论，但将其翻译成更实际的受众仍存在差距。 在本信件中，作为 Renson 等的补充，我们提供了如何构建 TMLE 的更易理解的说明。 显示英文摘要

摘要： Use of machine learning to estimate nuisance functions (e.g. outcomes models, propensity score models) in estimators used in causal inference is increasingly common, as it can mitigate bias due to model misspecification. However, it can be challenging to achieve valid inference (e.g., estimate valid confidence intervals). The efficient influence function (EIF) provides a recipe to go from a statistical estimand relevant to our causal question, to an estimator that can validly incorporate machine learning. Our companion paper, Renson et al. 2025 (arXiv:2502.05363), provides a thorough but approachable description of the EIF, along with a guide through the steps to go from a unique statistical estimand to development of one type of EIF-based estimator, the so-called one-step estimator. Another commonly used estimator based on the EIF is the targeted maximum likelihood/minimum loss estimator (TMLE). Construction of TMLEs is well-discussed in the statistical literature, but there remains a gap in translation to a more applied audience. In this letter, which supplements Renson et al., we provide a more accessible illustration of how to construct a TMLE.