Local Constraint-Based Causal Discovery under Selection Bias

TL;DR

This paper introduces a new method based on Y-Structure patterns for causal discovery under selection bias, demonstrating its effectiveness through simulations and real-world microarray data analysis.

Contribution

The paper develops a sound, finite-sample scoring rule for Y-Structures that predicts causal relations under selection bias, including cycles, and extends causal discovery methods.

Findings

Y-Structure patterns are sound for causal prediction under selection bias.

The scoring rule accurately predicts causal relations in simulations.

The method performs well on real microarray datasets, reducing spurious correlations.

Abstract

We consider the problem of discovering causal relations from independence constraints selection bias in addition to confounding is present. While the seminal FCI algorithm is sound and complete in this setup, no criterion for the causal interpretation of its output under selection bias is presently known. We focus instead on local patterns of independence relations, where we find no sound method for only three variable that can include background knowledge. Y-Structure patterns are shown to be sound in predicting causal relations from data under selection bias, where cycles may be present. We introduce a finite-sample scoring rule for Y-Structures that is shown to successfully predict causal relations in simulation experiments that include selection mechanisms. On real-world microarray data, we show that a Y-Structure variant performs well across different datasets, potentially circumventing spurious correlations due to selection bias.