Comparison between association rule data mining and causal discovery

Association rule data mining and causal discovery are two important data processing methods, but the comparison research between them is scarce now. Based on the description and analysis of the characteristics of association rules and causal rules, this work compares them theoretically in three aspects: directivity, guidance to the human behaviors, deduction and induction between them. The result shows that the intrinsic mechanic relationships between things can be obtained by the causal discovery, then we can predict the association rules based on it. Finally, this two kinds of data mining methods are applied to a real census income data set. The compared mining result validates the anterior analysis result.