Generating Representative from Clusters of Association Rules on Numeric Attributes

Association rule is useful to describe knowledge and information extracted from databases. However, a large number of association rules may be extracted. It is difficult for users to understand them. It is reasonable to sum up the rules into a smaller number of rules called representative rules. In this papar, we applied a clustering method to cluster association rules on numeric attribute and proposed an algorithm to generate representative rules from the clusters. We applied our approach to a real database, adult database. As the result, we obtained 124 rules divided into 3 clusters. We compared the rule generating method with another rule selecting method.