The Probability of Causal Explanation

We present a probabilistic theory of causal explanations, which integrates probabilistic and causal knowledge. Unlike most other approaches where a causal explanation is a hypothesis that one or more causative events occurred , we deene an explanation of a set of observations to be the occurrence of a chain of causation events. These causation events constitute a scenario where all the observations are true. The underlying causal model enables us to compute probabilities of the scenarios from the conditional probabilities of the causation events. The notion of causation event, which was rst introduced in Peng and Reggia, 1987] and was claimed to be \the crucial innovation," was nonetheless underspeciied. We provide a more adequate deenition here and explain its relationship to co-occurrence. Although probabilistic causal inference is NP-hard in general, our algorithm exploits characteristics of admissible input to achieve eecient computation.