Minimal Association Rules and Mining Algorithm

Conventional mining algorithms often produce too many rules for decision makers to digest.Instead,the concept of minimal association rules is introduced from the aspect of application in this paper.Minimal rule set,which includes rules with single item as consequent and the minimal number of items as the antecedent,can be used to derive the same decisions as other association rules without information loss,while the number of minimal rules is much less than of all rules.A mining algorithm without redundant rules is proposed and the mining efficiency is improved.