Data-driven causal path discovery without prior knowledge - a benchmark study
Causal discovery broadens the inference possibilities, as correlation does not inform about the relationship direction. The common approaches were proposed for cases in which prior knowledge is desired when the impact of a treatment/intervention variable is discovered or to analyze time-related dependencies. In some practical applications, more universal techniques are needed and have already been presented. Therefore, the aim of the study was to assess the accuracies in determining causal paths in a dataset without taking into account the ground truth and the contextual information. This benchmark was performed on the database with cause-effect pairs, using a framework consisting of generalized correlations (GC), kernel regression gradients and absolute residuals criteria (AR), along with causal additive modeling (CAM). The best overall accuracy, 80 combination of GC, AR, and CAM. We also used proposed bootstrap simulation to establish the probability of correct causal path determination and for which pairs the inference appears indeterminate. In this way, the mean accuracy was improved to 83 used for preliminary dependence assessment, as an initial step for commonly used causality assessment frameworks or for comparison with prior assumptions.
READ FULL TEXT