基于滑动窗口的数据流高效用模糊项集挖掘

High Utility Fuzzy Itemsets Mining Over Data Stream Based on Sliding Window Model

高效用项集挖掘可以提供有趣的结果集,但并不能提供单个项的数量,因此,本文提出了高效用模糊项集.但是,现实世界的数据是不断出现的,需要实时处理新到来的数据.为解决当前高效用模糊项集不能处理数据流的问题,又提出了模糊效用列表(fuzzy utility list,FUL)结构用于存储当前窗口中项的批次号、项在事务中的事务标识符、项的模糊效用以及项的剩余模糊效用,该结构能有效的对批次进行插入和删除操作.最后,基于FUL提出了数据流高效用模糊项集挖掘算法.对真实数据集和合成数据集进行了广泛的实验,结果证实了算法的效率及可行性.

High-utility itemsets mining(HUI)can provide interesting itemsets,but cannot provide information on the number of items.Therefore,high utility fuzzy itemsets are proposed.However,real-world data is constantly emerging.Thus,new incoming data needs to be processed in real time.To solve the problem that the current high utility fuzzy itemsets cannot handle the data stream,a fuzzy utility list(FUL)structure is proposed to store the information of items,including batch number of items,the transaction identifier of the items,the fuzzy utility of items,and the reminding fuzzy utility of items.FUL can effectively insert and delete batches.Finally,based on FUL,a high utility fuzzy itemset mining algorithm on data stream is proposed,extensive experiments on real and synthetic datasets show the efficiency and feasibility of the algorithm.