基于相关系数的正、负关联规则挖掘算法

A algorithm for discovering positive&negative association rules based on correlation coefficient

负关联规则描述的是项目之间的互斥关系,它与传统的正关联规则有着同样重要的作用。然而,大多规则挖掘算法只能挖掘正规则而忽略了负规则的挖掘。本文利用统计学中相关系数的理论,提出一个能同时挖掘正、负关联规则的算法,实验表明该算法是有效的。

Description of negative association rules （NAR） focuses on mutually exclusive correlations among items. They play important roles just as traditional positive association rules（PAR） do. However, most algorithms are only for discovering positive rules and lose negative rule. According to the correlation coefficient theory in statistics ,this paper presents a new algorithm for discovering both positive and negative association rules . Experiment results demonstrate the algorithm is efficient.