一个基于频繁项集的时态数据挖掘算法  

An Algorithm of Temporal Data Mining Based on Frequent Items

作　　者：李少年[1] 孟志青[1] 田媛[1] 

机构地区：[1]湘潭大学信息工程学院计算机系,湖南湘潭411105

出　　处：《湘潭大学自然科学学报》2003年第3期21-25,共5页Natural Science Journal of Xiangtan University

摘　　要：研究了基于频繁项集的一个时态数据挖掘算法,首先,引进了基于频繁项集关联规则的概念,性质,然后,给出了基于时态约束关联规则的相关概念的定义和性质分析。最后,给出了在时态数据库中挖掘具有时态约束的最大频繁项集,并在此频繁项集生成时态关联规则的算法.由此导出了一种具有一定意义的时态关联规则挖掘方法,这种方法可以用于商品销售,股票价格等问题的知识发现,最后讨论了该算法在股票数据分析上的一个应用。This paper discusses a new problem of knowledge discovery, an algorithm of temporal data mining based on frequenl items. First, we introduce the conception and some properties of association rule based on frequent items. Then, the definition of association rule based on temporal constraint is given. Meanwhile the interrelate properties are analyzed. We give (he algorithm to discover the longest frequent items with temporal constraint and generate temporal association rule in terms of them, which imply the temporal association rule mining technique of some practical meaning. This technique can be used to discovery knowledge in fields such as stock price, commodity sale and so on. Finally, we discuss the application of this algorithm to stock data.

关 键 词：知泌发现 效据挖掘 频繁项集 时态型 时态关联规则 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]