Parametric algorithms for mining share frequent itemsets

Itemset share, the fraction of some numerical total contributed by items wh en they occur in itemsets, has been proposed as a measure of the importance of itemsets in association rule mining. The IAB and CAC algorithms are abl e to find share frequent itemsets that have infrequent subsets. These algor ithms perform well, but they do not always find all possible share frequent itemsets. In this paper, we describe the incorporation of a threshold fact or into these algorithms. The threshold factor can be used to increase the number of frequent itemsets found at a cost of an increase in the number of infrequent itemsets examined. The modified algorithms are tested on a larg e commercial database. Their behavior is examined using principles of class ifier evaluation from machine learning.