基于频繁模式树的正负项目集挖掘

传统的基于支持度—置信度框架的关联规则挖掘方法可能会产生大量不相关的、甚至是误导的关联规则,同时也不能区分正负关联规则。在充分考虑用户感兴趣模式的基础上,采用一阶谓词逻辑作为用户感兴趣的背景知识表示技术,提出了一种基于背景知识的包含正负项目集的频繁模式树,给出了针对正负项目集的约束频繁模式树的构造算法NCFP-Construct,从而提高了关联规则挖掘的效率和针对性,实验结果显示该方法是有效的。

The Mining of the Positive and Negative Item Sets Based on Constrain Frequent Tree

Traditional association rule mining method based on the support-confidence framework may produce a large number of irrelevant,even misleading rules,and can not distinguish the positive association rules from the negative ones.In this paper,on the basis of taking full account of user-interested model,using the first-order predicate logic to describe background knowledge interested by users,a novel constrained frequent pattern tree based on the background knowledge is presented which includes positive and negative item sets,and the construction algorithm NCFP-construct of constraint frequent pattern tree including negative item set is given,so that the pertinence and efficiency of association rules mining result is improved.In the end,the experimental results show that the method is effective.