一种有效的基于约束的关联规则发现算法  

Effective Algorithm for Mining Constrained Association Rules

作　　者：杨文杰[1] 胡明昊[1] 唐振民[1] 杨静宇[1] 

机构地区：[1]南京理工大学计算机科学与技术系,江苏南京210094

出　　处：《南京理工大学学报》2005年第1期109-112,共4页Journal of Nanjing University of Science and Technology

摘　　要：基于约束的关联规则挖掘是针对特定约束的规则的挖掘,挖掘的结果有着更好的针对性和实用性,Separate算法是现有的效果较好的算法,但有 2点不足:未修剪生成的候选集和候选项重复生成。对此该文提出了改进的SeparateP算法,算法中加入了对候选集的修剪,并且利用了项集重复生成的信息,使候选集的修剪更加有效快捷。实验表明,改进算法显著提高了原算法的效率。Mining constrained association rules mines some special constrained rules and the results are more pertaining and practical. Separate Algorithm is a good algorithm to mine constrained association rules and has two main shortages. One is that it scans candidate itemsets without pruning; the other is that it creates candidate items redundantly. An improved algorithm, Separate P, is presented. The pruning on candidate sets is added to the algorithm, and Separate P makes full use of the information that the itemsets create repeatedly. The pruning is made more efficient and faster. The experimental results show that Separate P can increase the Separate efficiency.

关 键 词：数据挖掘 关联规则 项约束 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]