The Importance of Being Positive in Causal Statistical Fault Localization: Important Properties of Baah et al.´s CSFL Regression Model

This paper investigates the performance of Baah et al.´s causal regression model for fault localization when an important precondition for causal inference, called positivity, is violated. Two kinds of positivity violations are considered: structural and random ones. We prove that random, but not structural nonpositivity may harm the performance of Baah et al.´s causal estimator. To address the problem of random nonpositivity, we propose a modification to the way suspiciousness scores are assigned. Empirical results are presented that indicate it improves the performance of Baah et al.´s technique. We also present a probabilistic characterization of Baah et al.´s estimator, which provides a more efficient way to compute it.