Conditional Counterfactual Causal Effect for Individual Attribution

Identifying causes of an event, also termed as causal attribution, is a commonly encountered task in many applications. Existing methods, mostly in the Bayesian or causal inference literature, suffer from two main drawbacks: 1) they cannot attribute for individuals, and 2) only attribute one single cause at a time and cannot deal with the interaction effect among multiple causes. In this paper, we propose an individual causal attribution method, called conditional counterfactual causal effect (CCCE), which utilizes the individual observation as the evidence and is able to handle the common influence and interaction effects of multiple causes simultaneously. We discuss the identifiability of CCCE and then provide the identification formulas under certain proper assumptions. Finally, we conduct experiments on simulated and real data to illustrate the effectiveness of CCCE, compared with many state-of-the-art attribution methods. © UAI 2023. All rights reserved.