基于间接关联规则的数据挖掘算法研究

简单数据集可以通过关联规则得到在数据间的相互关系；相当多的情况下，由于不能从关联规则得到隐藏在数据间的相互关系，需要按间接关联规则分析出数据项集在交易集合中出现的频度，挖掘隐藏在数据间的相互关系。文中通过使用概念分层和基于近邻的方法，探讨利用FP树产生的频繁项集，对候选关联检验其是否满足项对支持度条件，并利用这个频繁项集挖掘事务的间接关联，找到挖掘事务的间接关联的内在规律，构造出不依赖中介条件的间接关联挖掘算法。

Research of Data Mining Algorithms Based on Indirect Association Rules

Simple data set can get its correlation among data by the association rules. While in many other cases,the correlation hidden in the data cannot be found in this way. So,indirect association rules are needed to analyze database frequentncss in transaction set and reveal the correlation hidden in the data. The research uses the concept hierarchy and affinity methods to discuss the frequent item sets occurred in the FP-tree, and examine the candidate association test to verify whether it is eligible for the support condition. Then use the frequent item sets to find the indirect association and its inherent rules in order to build indirect association mining algorithm free of intermediary conditions.