关联规则的鲁棒性分析与应用

Analysis and application on robustness of association rules

首先对采用支持度、置信度和改善度阈值来选择关联规则的不足进行了分析,然后引入规则的鲁棒性分析,提出了基于lift条件下的关联规则鲁棒性判断,能够在若干规则中选择出鲁棒性高的规则,提供决策参考,最后给出了实例应用。

Some shortages of association rules mining based on support-confidence and lift calculation are analyzed.Then the robustness of association rules are introduced to judge the interestingness of rules.The paper proposes a measure based on the robustness of association rules,which can be used to select high robustness rules for decision aiding.Finally an application is presented.