基于兴趣度的关联规则挖掘研究

Research on Association Rule Mining Based on the Interestingness

关联规则挖掘是数据挖掘领域的一个重要研究方向,规则膨胀的问题是困扰规则发现和应用的主要问题。为了得到数量少、质量优的关联规则,给出了一种同时挖掘正、负有趣关联规则的算法。采用置信度增量替代置信度,引用一种兴趣度量标准,避免生成相互矛盾的、虚假的关联规则。在得到关联规则后,又对预期的和可推导的关联规则进行了修剪。

Association rules mining is an important research direction in data mining field.Rules explo sion is a key problem in rules discovery and rules application.In order to obtain a rule set with small in quantity and excellent in quality,an algorithm mining both positive and negative interesting association rules is proposed.In the algorithm,no contradictory and false association rules are generated by substitut ing confidence with confidence increment and citing an interesting measure.Moreover,the expectable and derivable rules in the discovered rule set are pruned.