【摘要】 传统关联规则挖掘算法所讨论的只是一次事务内部的模式,但实际应用中随着时间的推移,事务之间会有某种联系或发展趋势,传统关联规则算法则无法揭示其中的规律。将在SPAM算法的基础上进行改进,加入通用时间约束,提出一个新算法——TSPAM算法。TSPAM算法采纳SPAM算法中的相关概念和方法,在SPAM算法的裁减部分和候选集的生成部分进行改进,对其加入时间约束。实验表明新算法对挖掘具有时间特征的序列模式是行之有效的。更多还原

【Abstract】 What the traditional association rules discuss is just a pattern of inside once trade,however,in the practical application by the time on,there will be some relations or developing trend among trades and the traditional association rules can not discover this law.This paper proposes a new Algorithm-TSPAM Algorithm.It accepts the ordered sequential tree,deep priority searching,and data expressing way etc. related concepts and methods,improves the curtail part and the producing waiting chosen items part,and adds the timing constraints in it.Experimental results demonstrate that algorithm is fairly effective for solving the sequential pattern mining problem on timing constraints.更多还原