多属性约束事件序列的关联规则挖掘方法

Mining multi-attribute event Sequential Pattern based on association rule

传统序列模式挖掘算法往往忽略了序列模式本身的时间特性,所考查的序列项都是单一事件,无属性约束。提出了一种挖掘多属性约束事件序列关联规则的方法。此方法基于传统的Apriori和AprioriAll算法,考虑了应用环境下事件序列模式中事件之间的过渡时间,采用分层式挖掘思想,先挖掘频繁序列模式,然后从频繁事件序列中挖掘多属性约束项的关联规则。实例分析为挖掘带时间限多属性约束的序列模式提供了实施思路。

The time trait is often ignored in the course of mining traditional sequential pattern, in which the sequential item is also without attribute constraint. An idea is given to mine event sequential pattern with muhi-attribute constraint in this paper. Based on the algorithm of Apriori and Apriori, the transition time is taken into account between events. According to the layer idea, the key task is to mine the frequent sequential pattern first, then to find out the association rules in the attribute constraint item. In the end it provides a way to mine sequential pattern with multi-attribute constraint by example analysis.