减少候选项集的数据流高效用项集挖掘算法
发布时间：2018-10-08 17:46
【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。
[Abstract]:In big data environment, too many candidate itemsets in the mining algorithm of high-utility itemsets greatly reduce the space-time efficiency of the algorithm. Therefore, a data stream high-utility itemset mining algorithm for reducing candidate itemsets is proposed. A global tree is established by scanning the current window in the data stream, and the redundant utility value of the header table entry and node in the global tree is reduced. The candidate pattern is generated based on the global tree and the candidate set utility of the local tree is reduced based on the growth algorithm, and the high utility pattern is selected from the candidate pattern. Experimental results based on real data flow show that the space-time efficiency and memory occupancy ratio of the algorithm are better than those of other data streams.
【作者单位】： 新乡学院计算机与信息工程学院;河南大学计算机与信息工程学院;
【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)
【分类号】：TP311.13
[Abstract]:In big data environment, too many candidate itemsets in the mining algorithm of high-utility itemsets greatly reduce the space-time efficiency of the algorithm. Therefore, a data stream high-utility itemset mining algorithm for reducing candidate itemsets is proposed. A global tree is established by scanning the current window in the data stream, and the redundant utility value of the header table entry and node in the global tree is reduced. The candidate pattern is generated based on the global tree and the candidate set utility of the local tree is reduced based on the growth algorithm, and the high utility pattern is selected from the candidate pattern. Experimental results based on real data flow show that the space-time efficiency and memory occupancy ratio of the algorithm are better than those of other data streams.
【作者单位】： 新乡学院计算机与信息工程学院;河南大学计算机与信息工程学院;
【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)
【分类号】：TP311.13
【相似文献】
相关期刊论文 前10条
1 李重周;杨君锐;;一种不产生候选项集的关联规则挖掘算法[J];重庆工学院学报(自然科学版);2007年06期
2 王云岚,李增智,屈科文;基于候选项集个数上阶的增量式关联规则更新算法[J];电子学报;2004年05期
3 刘晓玲;李玉忱;;一种不产生候选项集的关联规则挖掘算法[J];山东师范大学学报(自然科学版);2006年01期
4 陆楠,王U，
本文编号：2257698
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2257698.html
