LILI clustering algorithm: Limit Inferior Leaf Interval Integrated into Causal Forest for Causal Interference

2025-07-03

AI Summary

Causal forests suffer from systematic bias in average treatment effect (ATE) estimation because individual trees model treatment effects independently, leading to repeated classification errors across trees. To address this, we propose the Limit Inferior Leaf Interval (LILI) clustering algorithm—the first to introduce the set-theoretic concept of limit inferior into causal inference. LILI quantifies interval-wise similarity of covariates across leaf nodes of different trees, thereby establishing inter-tree dependencies and identifying reliable counterfactual pairs. By integrating leaf-level interval clustering with counterfactual matching, LILI achieves provably consistent ATE estimation while significantly improving accuracy and robustness. We establish theoretical consistency of the estimator under standard regularity conditions. Extensive experiments on multiple benchmark datasets demonstrate that LILI consistently outperforms state-of-the-art causal forests and matching-based methods.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Causal forest methods are powerful tools in causal inference. Similar to traditional random forest in machine learning, causal forest independently considers each causal tree. However, this independence consideration increases the likelihood that classification errors in one tree are repeated in others, potentially leading to significant bias in causal e ect estimation. In this paper, we propose a novel approach that establishes connections between causal trees through the Limit Inferior Leaf Interval (LILI) clustering algorithm. LILIs are constructed based on the leaves of all causal trees, emphasizing the similarity of dataset confounders. When two instances with di erent treatments are grouped into the same leaf across a su cient number of causal trees, they are treated as counterfactual outcomes of each other. Through this clustering mechanism, LILI clustering reduces bias present in traditional causal tree methods and enhances the prediction accuracy for the average treatment e ect (ATE). By integrating LILIs into a causal forest, we develop an e cient causal inference method. Moreover, we explore several key properties of LILI by relating it to the concepts of limit inferior and limit superior in the set theory. Theoretical analysis rigorously proves the convergence of the estimated ATE using LILI clustering. Empirically, extensive comparative experiments demonstrate the superior performance of LILI clustering.

Problem

• Reduces bias in causal effect estimation via LILI clustering
• Improves prediction accuracy for average treatment effect (ATE)
• Connects causal trees using confounder similarity in datasets

Innovation

• Integrates LILI clustering into causal forest
• Links causal trees via confounder similarity
• Reduces bias and improves ATE accuracy