Moment+: Mining Closed Frequent Itemsets over Data Stream

摘要

Closed frequent itemsets(CFI) mining uses less memory to store the entire information of frequent itemsets thus is much suitable for mining stream. In this paper, we discuss recent CFI mining methodsrnover stream and presents an improved algorithm Moment+ based on the existent one Moment. Moment+ focuses on the problem of mining CFIrnover data stream sliding window and proposes a new structure ExtendedrnClosed Enumeration Tree(ECET) to store the CFIs and nodes' BPNrnwhich is introduced to reduce the search space, with which new miningrnmethod is designed to mine more rapidly with little memory cost sacrifice. The experimental results show that this method is effective and efficient.