Re-examining Granger Causality with Causal Bayesian Networks and Reichenbachs Principles
Researchers have developed a new algorithm called causalized Granger causality (c-GC) to provide a more rigorous causal interpretation for Granger causality (GC). This updated method reinterprets GC using causal Bayesian networks and Reichenbach's principles, addressing criticisms about GC's lack of a strong causal foundation. The c-GC algorithm has demonstrated theoretical and graphical validity, showing promising results on synthetic data for causal discovery in observational datasets. AI
IMPACT Enhances causal inference methods applicable to AI models trained on time-series data.