Peculiarity Oriented Multi-database Mining

The paper proposes a way of mining peculiarity rules from multiply statistical and transaction databases. We introduce the peculiarity rules as a new type of association rules, which can be discovered from a relatively small number of the peculiar data by searching the relevance among the peculiar data. We argue that the peculiarity rules represent a typically unexpected, interesting regularity hidden in statistical and transaction databases. We describe how to mine the peculiarity rules in the multi-database environment and how to use the RVER (Reverse Variant Entity-Relationship) model to represent the result of multi-database mining. Our approach is based on the database reverse engineering methodology and granular computing techniques.