互斥关系模式挖掘算法研究

Study on algorithm for mining exclusive relation patterns

序列模式挖掘是数据挖掘的一个重要领域,结构关系模式挖掘是在序列模式挖掘基础上提出的一种新的挖掘任务。重点对结构关系模式的一个重要分支——互斥关系模式进行了研究,在给出与互斥关系模式相关概念的基础上讨论了互斥关系模式挖掘的两种算法,即基本检测法和分类检测法。实验结果表明,两种算法都是有效的,在序列模式数量很大时,分类检测法的挖掘效率高于基本检测法。结构关系模式挖掘和序列模式挖掘一样在实际应用中有着重要的价值,一些在序列模式挖掘过程中不能发现的隐藏模式将在结构关系模式中被发现,互斥关系模式的研究将进一步为结构关系模式挖掘理论的完善提供支持。

Sequential patterns mining is an important area of data mining. Structural relation pattern mmmg is a new Kind of data mining task which is proposed based on sequential pattern mining. Exclusive relation pattern is one of important forms of structural relation pattern, based on research of some definitions related to exclusive relation patterns, two algorithms for mining exclusive relation pattern are proposed, one is basic check method, the other is classification method. The experimental results indicate that the two algorithms are useful, and when the number of the sequential patterns is lager, classification method is more efficient. The structure relation pattern mining is very valuable in practical applications same as sequential Pattern mining. Some conclusions which cannot be given by sequential pattern mining is obtained through the structure relation pattern mining, and some hidden relations among sequential pattern is found. The further study on exclusive relation patterns will provide support for the perfect of structural relation patterns theory.