Directed Cyclic Graphs for Simultaneous Discovery of Time-Lagged and Instantaneous Causality from Time-Series Data

Abstract

We consider the problem of causal discovery from time-series observational data. We develop a novel framework that simultaneously discovers the time-lagged causality and the possibly cyclic instantaneous causality. Under common causal discovery assumptions, we prove the proposed model is generally identifiable by exploiting the instrumental information commonly available in time-series data. To the best of our knowledge, this is the first causal identification theory for directed graphs with general cyclic patterns that achieves unique causal identifiability. Structural learning is carried out in a fully Bayesian fashion. Through extensive simulations and an application to the Women’s Interagency HIV Study, we demonstrate the identifiability, utility, and superiority of the proposed model against state-of-the-art alternatives for time-series causal discovery.