基于效用模式树的高效用频繁模式挖掘算法  

Mining algorithm for high utility frequent patterns based on utility pattern tree

作　　者：王晨[1,2] 杜义华[1] 卢景秀[1] 李新[1] 

机构地区：[1]中国科学院计算机网络信息中心,北京100190 [2]中国科学院大学,北京100049

出　　处：《计算机应用》2013年第A02期111-115,141,共6页journal of Computer Applications

摘　　要：为了提高效用模式挖掘的效率,提出了一种基于效用模式树的两阶段效用模式挖掘算法。在第一阶段,该算法能够对全局非候选节点进行合理的舍弃,并能降低全局效用模式树中节点的估计效用。通过效用模式增长算法,递归地生成候选高效用项集。在第二阶段,通过扫描调整后的事务数据库,缩小第二阶段的搜索空间。实验表明,该算法能够减少候选集的数量,提高高效用项集的生成效率。In order to improve the efficiency of utility pattern mining, a two-phase, utility tree based mining algorithm was proposed. In phase one, reasonable non-candidate node discarding was implemented on utility pattern tree, which could significantly reduce estimated node utility of global utility pattern tree. Through utility pattern growth algorithm, candidate itemsets generated recumively. In phase two, by scanning the adjusted transaction database, the search space was reduced. Experimental results demonstrate that the algorithm can reduce the number of candidate sets, and improve the efficiency of the generation about high utility itemsets.

关 键 词：效用挖掘 频繁模式 效用模式树 估计效用 数据挖掘 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]