Top-k高模糊效用项集挖掘算法  

Top-k high fuzzy utility itemsets mining algorithm

作　　者：王斌 周伟 李晓华 胡克勇 WANG Bin;ZHOU Wei;LI Xiao-hua;HU Ke-yong(School of Information and Control Engineering,Qingdao University of Technology,Qingdao 266520,China)

机构地区：[1]青岛理工大学信息与控制工程学院,山东青岛266520

出　　处：《计算机工程与设计》2023年第11期3312-3319,共8页Computer Engineering and Design

基　　金：国家自然科学基金项目(61902205)。

摘　　要：针对高模糊效用项集挖掘算法中存在的阈值选择难题,提出一种Top-k高模糊效用项集挖掘算法TKHFU(top-k high fuzzy utility itemset mining)。将高模糊效用项集挖掘算法中设定阈值的问题转化为设定所需高模糊效用项集数量k的问题。提出模糊项集效用列表结构,有效保存事务数据库中的关键信息,避免项集间复杂的连接操作。提出一种更紧密的模糊效用上界和两种剪枝策略,将剪枝策略运用至列表中,减少搜索空间。实验结果表明,该算法在运行时间、内存消耗及可伸缩性方面均优于最新算法。Aiming at the problem of threshold selection in high fuzzy utility itemsets mining algorithm,TKHFU(Top-k high fuzzy utility itemsets mining)algorithm was proposed.The problem of setting the threshold was transformed into the problem of setting the required number k of high fuzzy utility itemset in mining high fuzzy utility itemset.Fuzzy itemset utility list structure was proposed,which efficiently saved the key information in the transactional database and avoided the complex itemsets joining operations.A tighter fuzzy utility upper bound and two pruning strategies were proposed,and the pruning strategies were applied to the list,reducing the searching space.Experimental results show that TKHFU outperforms the state-of-the-art algorithms in terms of running time,memory consumption and scalability.

关 键 词：数据挖掘 项集挖掘 阈值 高模糊效用 效用列表 模糊效用上界 剪枝策略 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]