Exploding number of frequent itemsets in the mining of negative association rules

This paper uses the over-frequent itemset concept to solve the problem of an exploding number of frequent itemset including negative items when mining negative association rules.The analysis shows that with the over-frequent itemset and all its interesting association rules cannot be generated and an upper bound to the negative items must be included in the itemset during the generation of frequent itemset.The paper brings an algorithm which restricts the generation of frequent itemset by using 3 parameters: min support,max support,and the maximum number of negative item in the itemset.Test results show that the latter two parameters are necessary.More specifically,when the data pool has a large number of items,the normal mining process can only be ensured by appropriately setting these two additional parameters.