摘要
结构关系模式挖掘是本课题组提出的一种新的数据挖掘理论,主要研究序列之间的并发关系、互斥关系、重复关系等.并发序列模式挖掘是结构关系模式挖掘的重要组成部分.文中从序列间的相对关系出发研究并发关系,给出并发序列模式的相关性质,并对现有并发序列模式挖掘算法进行优化.通过实验对比可以看出:该算法根据并发序列模式的反单调特性和非平凡特性,对挖掘结果进行大幅精简,使得挖掘更有实际意义.
Theory of Structure Relation Patterns mining was a new data mining theory which our research group proposed.It mainly researched the relations among sequences including concurrent relation,exclusive relation,iterative relation and so on.Concurrent sequential patterns mining was an important part of Structure Relation Patterns mining.In this article we considered the concurrent relation from the relative relationship among sequential patterns,and relevant properties were also given here.We optimized the existing algorithm used to mine concurrent sequential patterns.Through experiment,the mining result was severely curtailed contrast to the existing algorithm,and it made the concurrent sequential patterns mining more meaningful.
出处
《沈阳化工大学学报》
CAS
2011年第3期267-272,共6页
Journal of Shenyang University of Chemical Technology
基金
辽宁省教育厅科学研究计划资助项目(05L338)
关键词
并发关系
并发度
并发序列模式
并发序列模式挖掘
concurrent relation
concurrence threshold
concurrent sequential pattern
concurrent sequential patterns mining
