Using Closed Itemsets for Discovering Representative Association Rules

Abstract

A set of association rules is called representative if it is a minimal set of rules from which all association rules can be generated. The existing algorithms for generating representative association rules use all the frequent itemsets as input. We present a new approach for generating representative association rules that uses only a subset of the set of frequent itemsets called frequent closed itemsets. This results in a big reduction in the input size and, therefore, faster algorithms for generating representative association rules. Our approach uses ideas from formal concept analysis to find frequent closed itemsets.