基于共现结构的频繁高效用项集挖掘算法

Frequent and High-Utility Itemset Mining Algorithms Using Estimated Utility Co-Occurrence Pruning

频繁项集挖掘和高效用项集挖掘是数据挖掘研究中的重要内容.为克服在实际应用中单独使用这2类算法的局限性,频繁高效用项集挖掘算法开始被提出.基于经典高效用项集挖掘(Fast High-Utility Miner,FHM)算法,本文提出了频繁高效用项集挖掘(improved FHM with Support,iFHMS)算法.该算法构建了1个改进的共现结构(Utility and Support Co-occurrence Structure,USCS),存储满足效用约束条件和支持度约束条件的2-项集的事务加权效用和支持度.通过实验得出结论,iFHMS算法能够有效发现频繁高效用项集,且在时间效率方面有一定程度提升.

Frequent itemset mining and high-utility itemset mining are two important branches in data mining research.In order to overcome the limitations of using these two kinds of algorithms alone in practical applications,frequent and high-utility itemset mining algorithm has been proposed.In this paper,we propose a frequent and high-utility itemset mining algorithm,named iFHMS(improved FHM with Support).The algorithm constructs an improved Co-occurrence Structure to store TWU and support of 2-itemsets which are not less than thresholds.It can be concluded from the result of experiments that iFHMS can effectively find frequent and high-utility itemsets,and improve the time efficiency to a certain extent.