基于属性位复用的约束性关联规则挖掘算法

在提取满足用户特定需求的关联规则时,由于现有约束性关联规则挖掘算法存在大量的冗余候选项和重复计算,故提出一种基于属性位复用的约束性关联规则挖掘算法,其适合挖掘任何长度且满足用户特定需求的关联规则。该算法通过属性位的权值组合,将交易事务转换成整数,用属性位复用技术构建候选区间,并利用其端点值双向变化,构建索引候选频繁项,同时也用布尔运算计算其支持数。实验证明其比现有算法更快速,将其应用到客户关系管理系统中分析客户关联信息,可以有效地提高系统效率。

Constraint association rule mining algorithm based on attribute location multiplexing

When extracting association rule to meet specific demand given by user,as the existing constraint association rules mining algorithms have superfluous candidate and repeated computing.Constraint association rule mining algorithm based on attribute location multiplexing is proposed,which is suitable for mining any long association rule to meet specific demand given by user.The algorithm turns transaction into integer by weights combination of attribute location,and uses attribute location multiplexing to create candidate interval,and uses value of its endpoints to double vary to generate indexical candidate frequent item sets,and uses Boolean operation to compute support.This experiment indicates that the efficiency is faster than the existing algorithms.The algorithm fast improves the efficiency when it is applied to customer relationship management system to analyze custom association.