시간 변화 처리 효과 추정을 위한 종단적 효율적 조정 집합: 비모수적 인과 그래프 모델
본 연구는 시간 변화 처리 효과 추정을 위한 새로운 종단적 조정 집합 정의를 제안하고, 이를 통해 기존 방법보다 낮은 분산을 가지는 추정량을 도출할 수 있음을 보였다.
Efficient Adjustment Sets for Estimating Time-Varying Treatment Effects in Nonparametric Causal Graphical Models
Extending previous results, this paper proposes a novel definition of sufficient time-dependent adjustment sets that can yield estimators with lower asymptotic variance compared to existing methods, by exploiting conditional independencies in the causal graph.
時間依存事象の因果推論における一般化ネルソン-アーレン推定量
逆確率加重法を用いて、時間依存事象の因果推論における一般化ネルソン-アーレン推定量を提案し、その漸近的性質を導出した。
Causal Representation Learning: Optimizing Representations to Preserve Interventional Control
Causal representations that preserve interventional control over a target variable while minimizing information about the input variables.
관찰 데이터에서 혼란 요인과 탈락에 대한 순차적 조정을 통한 인과 효과의 회복 및 추론
관찰 데이터에서 혼란 요인과 선택 편향을 동시에 해결하기 위한 순차적 조정 기준(SAC)을 제시하고, 이를 활용한 표적 최소 손실 추정(TMLE) 기반의 순차적 회귀 추정량(TSR)을 개발하였다.
Recovering Causal Effects with Sequential Adjustment for Confounding and Attrition
The authors introduce the sequential adjustment criteria (SAC) to recover causal effects in the presence of confounding and attrition bias. They also propose a targeted sequential regression (TSR) estimator that is multiply robust under certain conditions.
Identifying Complete Data Distributions from Partially Observed Data Using Causal and Counterfactual Reasoning
The core message of this article is that missing data problems can be viewed as a form of causal inference, where the goal is to identify the complete data distribution from the observed data distribution by leveraging graphical representations and counterfactual reasoning.
Comprehensive Framework for Detecting and Measuring Confounding Effects Using Causal Mechanism Shifts
This work proposes a comprehensive framework for detecting and measuring confounding effects among variables, including separating observed and unobserved confounding, and assessing the relative strengths of confounding between different variable sets. The framework leverages data from multiple contexts where causal mechanisms of variables have shifted.
Causal Inference in Networks: A GNN-Driven Instrumental Variable Approach for Estimating Treatment Effects in the Presence of Hidden Confounders
A novel GNN-driven instrumental variable approach, CgNN, that leverages network structure to mitigate hidden confounder bias and accurately estimate main, peer, and total causal effects in network data.
Incorporating Expert Knowledge to Restrict the Markov Equivalence Class of Maximal Ancestral Graphs
The core message of this work is to develop methods for incorporating expert or background knowledge about causal relationships into the analysis of maximal ancestral graphs (MAGs) in order to restrict the Markov equivalence class and improve causal identification.
© 2024 by Linnk AI