一类时态近似周期关联规则的知识发现问题

Knowledge discovery for a kind of association rules of temporal approximate periodicity

研究一类基于时态约束的属性状态之间存在关联的近似周期知识发现问题。首先构造了时态近似周期关联规则模型,然后提出了一个基于SOM(自组织特征映射网络)聚类来寻找近似周期关联规则的算法,对十多年来的股票数据和高频股票数据分别进行了一些实验,实验表明该算法是有效的。

This paper discusses a kind of association rules of temporal approximate periodicity based on temporal constraint. First it presents an association rules of the approximate periodicity pattern.Then,it discusses an algorithm based on self-organizing map to find approximate periodic association rules.Experiment results show that proposed algorithms are efficient.