A Critique of Inductive Causation

: In this paper we consider the problem of inducing causal relations from statistical data. Although it is well known that a correlation does not justify the claim of a causal relation between two measures, the question seems not to be settled. Research in the field of Bayesian networks revived an approach suggested in [16]. It is based on the idea that there are relationships between the causal structure of a domain and its corresponding probability distribution, which could be exploited to infer at least part of the causal structure from a set of dependence and independence statements. This idea was developed into the inductive causation algorithm [14]. We review this algorithm and examine the assumptions underlying it.

• Type: Conference
• Year: 1999
• Where: ECSQARU
• Authors: Christian Borgelt, Rudolf Kruse
• Keywords: Automated Reasoning, Causal Relation, Causal Structure, Correlation Does Not, ECSQARU 1999