摘要
研究分布式环境下约束性关联规则更新问题,包括数据库中事务增加和删除2种情况.引入向导集的概念,提出基于全局局部模式的约束性关联规则增量式更新算法DUCAR,其中包括局部约束性频繁项目集更新算法ULFC和全局约束性频繁项目集更新算法UGFC.该算法充分利用原先的挖掘结果提高更新效率,首先从最高维的频繁n项目集进行更新,在更新过程中考虑约束条件,结合剪枝算法,生成较少数量的满足约束条件的候选项目集.将该算法用Java加以实现,采用多组数据对此算法的性能进行测试,并与其他算法作对比实验,实验结果表明,该算法是高效可行的.
A fast incremental updating technique is presented for maintaining the constrained association rules discovered in the cases including insertion and deletion of transactions in the distributed databases. The concept of induced set is introduced. The efficient algorithm DUCAR ( distributed upda- ting of constrained association rules), which includes algorithms ULFC (updating of local frequent constrained itemsets) and UGFC (updating of global frequent constrained-itemsets), is proposed to update constrained association rules in distributed databases. The algorithm makes full use of the previous mining result to cut down the cost of updating frequent itemsets with item constraints in the distributed databases. The algorithm starts by computing the highest n level frequent itemsets in the original databases, and generate a small number of candidate itemsets by exploring pruning technique. Finally, algorithm DUCAR is implemented by Java and a group of dataset is applied to test the performance of algorithm and the experiment result is compared with other algorithm. The experiment results show that the algorithm is effective and efficient.
出处
《东南大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2006年第1期34-38,共5页
Journal of Southeast University:Natural Science Edition
基金
国家自然科学基金资助项目(70371015)
江苏省重点实验室开放基金资助项目(KJS03064)
关键词
关联规则
项约束
约束性频繁项目集
频繁项目集更新
分布式数据挖掘
association rule
item constraints
frequent itemsets with item constraints
frequent itemsets updating
distributed data mining