动态关联规则的趋势度挖掘方法  

Method of data tendency measure mining in dynamic association rules

作　　者：张忠林[1] 曾庆飞[1] 许凡[1] 

机构地区：[1]兰州交通大学电子与信息工程学院,兰州730070

出　　处：《计算机应用》2012年第1期196-198,共3页journal of Computer Applications

基　　金：国家自然科学基金资助项目(61163010);甘肃省科技支撑计划项目(1011GKCA040)

摘　　要：针对规则随着时间变化的特点,在分析原有定义和对支持度向量(SV)和置信度向量分类的基础上,提出了动态关联规则趋势度的挖掘方法。首先,利用趋势度阈值消除无价值的规则,减小候选项集;其次,产生动态关联规则的趋势度元规则,找出具有价值的规则,提高挖掘质量;最后,通过对具有增减和周期趋势的事物数据库分析,证明了所提方法的有效性。Based on the original definition and classification of Support Vector（SV） and confidence vector,this paper put forward a method of data tendency measure mining in dynamic association rules,according to the characteristic of rules with time changing.First,taking advantage of tendency measure threshold to eliminate useless rules,the item sets candidates can be reduced.Second,producing the dynamic association rule,this method found out valuable rules and improved the mining quality.Finally,by analyzing a transaction database that is characterized by the tendency of changes and cycles,the analytical results verify the validity of the proposed method.

关 键 词：数据挖掘 动态关联规则 趋势度 元关联规则 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]