CBWon: a fast algorithm for activating remining of association rules

Nowadays, association rules mining has become one of the predominate tasks employed to discover informative rules from large data set to support decision-making. One of the major difficulties in applying associations mining technique is the setting of an appropriate minimum support. Unfortunately, a large support threshold would hinder the discovery of some rare but informative rules. In this paper, we propose a novel algorithm called CBW_on. By keeping and utilizing the set of frequent itemsets MF and an auxiliary set of infrequent α-itemsets MIFa the proposed CBW_on algorithm can significantly reduce, over an order of magnitude, the computation time spent on rediscovery of frequent itemsets.