摘要： 为缩减关联规则存储空间和方便查询关联规则，提出一种前件为单一项目的最小预测集算法。利用集合枚举树找到最大频繁项 目集，据此来挖掘最小预测集。对规则扩展的有效性进行证明。实验结果表明，通过该算法得到的最小预测集比传统方法小1个数量级。

关键词: 关联规则, 集合枚举树, 最小预测集, 最大频繁集

Abstract: For reducing the spaces of rule database and facilitating users to query, the minimal prediction set is used and mined using maximum frequent item sets which are found by a set-enumeration tree. The effectiveness of rule expansion is proved in theory. Experimental results show that it is efficient to reduce 1 order of the traditional one.

Key words: association rules, set-enumeration tree, minimal prediction set, maximum frequent item sets

中图分类号: 

• TP311