Discovering Association with Copula Entropy

Association (or dependence) is such a statistical tool defined for measuring the relationships between random variables of real systems [1]. Correlation, as the linear version of association, is the most commonly used one in real applications, while statistical dependence covers much broad types of associations including nonlinear cases than correlation does. Another closely related concept, Causality is defined for causal relationships in physical, social and biological systems. Even it is well known that association does not imply causation, association is still a necessary condition for causality in general. Association and causality are of significant importance in healthcare and medicine [1, 2]. In medical research, association is widely used as first evidence for scientific discoveries. Causality is much fundamental in all branches of medicine - clinicians diagnose based on symptom-disease relationships, pharmacologists find drugs according to drugs' effect on disease, epidemiology study how environmental factors affect Ma Jian is with Hitachi (China) Research & Development Corporation, Beijing 100084, China.