减少候选项集的数据流高效用项集挖掘算法  

High utility itemsets mining algorithm of data stream with reducing candidate itemsets

作　　者：茹蓓[1] 贺新征[2] 

机构地区：[1]新乡学院计算机与信息工程学院,河南新乡453003 [2]河南大学计算机与信息工程学院,河南开封475004

出　　处：《计算机应用研究》2017年第11期3379-3383,共5页Application Research of Computers

基　　金：河南省科技厅软科学研究计划资助项目(152400410345);河南省科技厅科技攻关资助项目(172102210445);河南省教育厅资助项目(15A520093)

摘　　要：大数据环境下高效用项集挖掘算法中过多的候选项集极大地降低了算法的时空效率,为此提出了一种减少候选项集的数据流高效用项集挖掘算法。通过数据流中当前窗口的一次扫描建立一个全局树,并降低全局树中头表入口与节点的冗余效用值。基于全局树生成候选模式,基于增长算法降低局部树的候选项集效用,从候选模式中选出高效用模式。基于真实数据流的实验结果表明,算法的时空效率与内存占用比均优于其他数据流的高效用模式挖掘算法。In the big data stream scenario, high utility pattern mining algorithm generated a lot of candidate itemsets and reduced the efficiency of time and space of algorithm. This paper proposed a high utility itemsets mining algorithm of data stream with reducing candidate itemsets to resolve that problem. Firstly, it constructed a global tree through a single scan of the current window in a data stream, reduced redundancy utilities in both entries of a header table and nodes in the tree in this stage. Secondly, it generated candidate patterns from the constructed tree, reduced the redundancy utilities of local tree by growth algorithm. Lastly, it identified a set of high utility patterns from the candidate patterns. Realistic data streams based experimental results show that the proposed algorithm performs better in efficiency of time and space and memory usage index than the other high utility pattern mining algorithm of data streams.

关 键 词：大数据 数据流 高效用项集 模式挖掘 模式增长 候选模式 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]