Causal Inference via Nonlinear Variable Decorrelation for Healthcare Applications

TL;DR

This paper introduces a novel method for causal inference in healthcare that decorrelates features under nonlinear conditions and employs association rules for interpretability, validated through extensive experiments and expert evaluation.

Contribution

The paper presents a new variable decorrelation regularizer for nonlinear confounding and uses association rule mining to enhance model interpretability in healthcare.

Findings

Superior performance in parameter estimation and causality computation.

Validated effectiveness and interpretability through expert evaluation.

Effective handling of nonlinear confounding in healthcare datasets.

Abstract

Causal inference and model interpretability research are gaining increasing attention, especially in the domains of healthcare and bioinformatics. Despite recent successes in this field, decorrelating features under nonlinear environments with human interpretable representations has not been adequately investigated. To address this issue, we introduce a novel method with a variable decorrelation regularizer to handle both linear and nonlinear confounding. Moreover, we employ association rules as new representations using association rule mining based on the original features to further proximate human decision patterns to increase model interpretability. Extensive experiments are conducted on four healthcare datasets (one synthetically generated and three real-world collections on different diseases). Quantitative results in comparison to baseline approaches on parameter estimation and causality computation indicate the model's superior performance. Furthermore, expert evaluation given by healthcare professionals validates the effectiveness and interpretability of the proposed model.