基于关联系数的Apriori算法改进

Updating of Apriori Algorithm Based on Correlation Coefficientr

分析了Apriori核心算法,说明了其设计思想上的不足,引进了兴趣度的概念。主要定义了基于协方差关联系数的兴趣度,并因此而设计了基于此兴趣度定义的关联规则挖掘算法,该算法能够更准确地挖掘出交易集中不同产品间的紧密相关性,减少产生的无用规则。

Based on the analysis of apriori algorithm, the paper points out the defects of its design and adopts the concept of interest measure, defines the interest measure that is based on the covariance correlation coefficient. Therefore the paper designs a mining algorithm of association rules which is based on the new definition of interest measure. The algorithm can more accurately mine the close relations among different products in the database and decrease the useless rules.