Fuzzy meta-association rules for information fusion

Nowadays, data volume, distribution, and volatility makes it difficult to apply traditional Data Mining techniques in the search of global patterns in a domain under observation. This is the case of the methods for discovering associations, which typically require a single uniform dataset. To address the scenarios in which satisfying this requirement is not practical or even feasible, we propose a new method for fusing information extracted from individual and partially heterogeneous databases in the form of association rules. This method produces meta-association rules; i.e., rules in which the antecedent or the consequent may contain rules as well. In this paper, we describe the formulation and the implementation of two alternative frameworks that obtain, respectively, crisp meta-rules and fuzzy meta-rules. The comparison of both frameworks shows that the fuzzy approach offers several advantages: it is more accurate, produces a more manageable set of rules for human inspection, and allows the incorporation of contextual information to the mining process expressed in a more human-friendly format.