基于决策属性的关联规则挖掘

针对具有决策属性的数据库模型,提出了高效挖掘关联规则算法,即矩阵划分算法:根据决策属性将扫描后的数据库划分成两个包含不同决策属性的矩阵,分别采用向量法挖掘频繁项目集.关联规则的生成可充分利用“与”运算的优点,查找规则前件或后件的支持度.所提出的算法减少了候选二项频集的生成,以及“与”运算的大小,与apriori算法及传统的向量法挖掘关联规则相比,效率明显提高.

Mining Based on Association Rules of Decision Property

This paper puts forward efficient algorithm according to the database model that has decision property to mine association rules.Algorithm of matrix partition:the database is divided into two matrixes that contain different property,and use vector to mine frequent itemsets.Producing of rules can make use of the advantage of "and" operation to find support of the antecedent or the consequent.This algorithm reduced the production of candidate two-itemsets and amount of operation.Compare to apriori and other traditional algorithm,the advantage of algorithm efficiency rises obviously.