Information flows in causal networks

Abstract

We use a notion of causal independence based on intervention, which is a fundamental concept of the theory of causal networks, to define a measure for the strength of a causal effect. We call this measure “information flow” and compare it with known information flow measures such as transfer entropy