一种含负项目的一般化关联规则挖掘算法

Algorithm of mining general association rules with negative items

传统的关联规则是形如A B反映正项目之间关联关系的蕴涵式,它无法反映出数据之间隐藏的负关联关系。在表达式中引入负项目,将这种传统的关联规则扩展成包含正、负项目的一般化关联规则。介绍了一般化关联规则的概念及其相关性质定理,并加以证明,提出了一种基于频繁模式树的挖掘混合正、负项目的一般化关联规则的MGPNFP算法,对其性能进行了分析,并比较了MGPNFP算法比现有的挖掘含负项目关联规则的算法所具有的优势。

The traditional association rule is an expression as A=Bwhich reflects the relation among positive items. But which can＇t reflect the negative association hidden in data. The negative items are introduced to expressions, and the traditional association rules are expanded to the general association rules with positive and negative items. The concept and qualities of general association rules is introduced, these theorems is proved, a MGPNFP algorithm of mining general association rules blending with positive and negative itcrns is proposed based on frequent pattern tree, and MGPNFP algorithm＇s predominance of performance compared with other algorithms of mining association rules＇ with negative items is analyzed.