基于项集支持度的关联规则增量更新算法——BISIUA  

Incremental updating algorithm based on support of itemset—BISIUA

作　　者：李志刚[1] 李峰[2] 张善姝[2] 王汝琳[1] 

机构地区：[1]中国矿业大学信息工程学院,北京100083 [2]河北理工大学计算机与自动控制学院,河北唐山063009

出　　处：《计算机工程与设计》2007年第17期4072-4074,4078,共4页Computer Engineering and Design

基　　金：唐山市科技局科技攻关计划基金项目(04204501C-23)

摘　　要：对当最小支持度和最小置信度都不变的情况下数据库中数据量增加时的关联规则增量更新问题进行了研究。给出了一个简单的判定公式,依据项集在原数据库DB和新添加的数据库db中的实际支持度来判定该项集在更新后的数据库DB∪db中是否频繁。对Apriori算法进行了改进,使其能适应本增量更新算法。对BISIUA算法思想进行了详细的论述,并在此基础上给出了该算法的形式化描述。由理论分析可知,BISIUA算法能有效提高关联规则增量更新的效率。The problem ofincremental updating is studied based on association roles when minsupport and minconfidence all keep invariable but the quantity of the data which in database keep increase. It has produced a simple determination formula, the itemset is or not frequent in the incremental database of DBUdb which lies on the assured support in the original database of DB and in the new increased database of db. To adapt this incremental updating algorithm, a few improvement is made to the Apriori algorithm. The elaboration of the BISIUA algorithm is discussed, and formalized description is shown for the algorithm. The BISIUA algorithm can effectively enhance the efficiency of the problem of incremental updating based on association rules according to the theoretical analysis.

关 键 词：数据挖掘 关联规则 增量更新 支持度 项集 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]