An Itemset-Driven Cluster-Oriented Approach to Extract Compact and Meaningful Sets of Association Rules