数据流频繁模式挖掘算法设计  

Designing Algorithm on Mining Frequent Patterns in Data Streams

作　　者：敖富江[1] 颜跃进[1] 黄健[1] 黄柯棣[1] 

机构地区：[1]国防科技大学机电工程与自动化学院,长沙410073

出　　处：《计算机科学》2008年第3期1-5,共5页Computer Science

基　　金：国家科学自然基金项目(项目编号:60573057)资助

摘　　要：介绍了数据流频繁模式的概念和定义,提出了数据流频繁模式挖掘算法的通用数据流处理模型,详细总结了数据流频繁模式挖掘算法的三种分类方式:"窗口模型"、"结果集类型"和"结果集精确性"。基于这些分类方法提出了数据流频繁模式挖掘算法的设计立方体,该立方体不仅涵盖了现有的数据流频繁模式挖掘算法,还对设计新的算法具有指导意义。基于设计立方体,分析了设计算法时应当采取的有效策略,旨在为设计新算法提供一个有力参考。最后讨论了数据流频繁模式挖掘的进一步研究工作。This paper introduces some concepts and definitions about frequent patterns in data streams, and presents the general data streams processing model for mining frequent patterns in data streams, and detailedly sums up three classifications of the algorithms on mining frequent patterns in data streams, including Window Model, the Type of Result Set and the Accuracy of Result Set. Based on these classifications, the cube for designing algorithm on mining fre- quent patterns in data streams is presented. The cube not only covers the exiting algorithms, but also allows for the design of new ones suited for various application requirements. Based on the cube, we analyse some valid strategies for designing the algorithm, aiming at presenting a powerful reference for designing new ones. Lastly, we discuss some future research works about mining frequent patterns in data streams.

关 键 词：数据流 频繁模式 设计立方体 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论] TN929.533[自动化与计算机技术—计算机科学与技术]