【作者】 张磊； 夏士雄； 周勇； 牛强；

【Author】 ZHANG Lei,XIA Shi-xiong,ZHOU Yong,NIU Qiang(School of Computer Science and Technology,China University of Mining and Technology,Xuzhou 221116,China)

【机构】 中国矿业大学计算机科学与技术学院；

【摘要】 现有的关联规则挖掘方法中,大多采用单一的最小支持度.实际上,应该根据数据的特点设置不同的最小支持度.文中针对这一问题,将语义信息引入关联规则挖掘之中,提出了具有语义最小支持度的关联规则挖掘方法.该方法首先计算项目之间的语义相关度,然后根据候选集的语义相关度对候选集合进行过滤,最后根据候选集的语义相关度,确定其语义最小支持度.实验表明:具有语义最小支持度的关联规则挖掘方法比传统的关联规则挖掘方法能够更好地实现关联规则的挖掘.更多还原

【Abstract】 The single minimum support degree is used in the existing association rules mining methods mostly.In fact,the different minimum support degrees should be set based on the characteristics of the data.Association rules mining method with semantic minimum support degree is proposed by importing semantics into association rules mining in the paper.Firstly,semantic relevance degree between the items is computed in the method.Secondly,the candidate sets is filtered according to their semantic relevance degree.Finally,the semantic minimum support is determined based on semantic relevance degree of the candidate set.Experiments showed that association rules mining method with semantic minimum support degree can mine association rules better than the traditional association rules mining method.更多还原