arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.27320stat.MEstat.ML
收藏

通过观察结果条件化来回顾性反事实预测:一种跨世界方法

Retrospective Counterfactual Prediction by Conditioning on the Factual Outcome: A Cross-World Approach

HEC Lausanne, University of Lausanne(洛桑大学高等商业研究学院)

Juraj Bodik

AI总结:

本文研究了在Neyman-Rubin超总体模型下,基于协变量x和观察到的结果y估计反事实结果μ(x,y)及构建有效预测区间的方法,通过跨世界相关性ρ(x)解决反事实未识别问题。

摘要:

回顾性因果问题询问如果个体接受不同治疗会发生什么。我们研究了估计μ(x,y)=E[Y(1)|X=x,Y(0)=y],即具有协变量x和观察到结果y的个体的预期反事实结果,并在Neyman-Rubin超总体模型下构建有效的预测区间。该量通常在没有额外假设的情况下无法识别。为了连接观察到和未观察到的潜在结果,我们使用跨世界相关性ρ(x)=cor(Y(1),Y(0)|X=x);合理界限上的ρ(x)使该问题得以解决。我们引入了回顾性反事实估计量\hatμ_ρ(x,y)和预测区间C_ρ(x,y),这些估计量在标准因果假设下渐近满足P[Y(1)∈C_ρ(x,y)|X=x,Y(0)=y]≥1-α。许多常见基线隐含对应于端点选择ρ=0或ρ=1(忽略事实结果或将反事实视为移位的事实结果)。通过跨世界依赖性在这些情况之间插值,在理论和实践中均获得显著收益。

英文摘要:

Retrospective causal questions ask what would have happened to an observed individual had they received a different treatment. We study the problem of estimating $μ(x,y)=\mathbb{E}[Y(1)\mid X=x,Y(0)=y]$, the expected counterfactual outcome for an individual with covariates $x$ and observed outcome $y$, and constructing valid prediction intervals under the Neyman-Rubin superpopulation model. This quantity is generally not identified without additional assumptions. To link the observed and unobserved potential outcomes, we work with a cross-world correlation $ρ(x)=cor(Y(1),Y(0)\mid X=x)$; plausible bounds on $ρ(x)$ enable a principled approach to this otherwise unidentified problem. We introduce retrospective counterfactual estimators $\hatμ_ρ(x,y)$ and prediction intervals $C_ρ(x,y)$ that asymptotically satisfy $P[Y(1)\in C_ρ(x,y)\mid X=x, Y(0)=y]\ge1-α$ under standard causal assumptions. Many common baselines implicitly correspond to endpoint choices $ρ=0$ or $ρ=1$ (ignoring the factual outcome or treating the counterfactual as a shifted factual outcome). Interpolating between these cases through cross-world dependence yields substantial gains in both theory and practice.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑