基于兴趣度的多值关联规则挖掘

Mining Quantitative Association Rules Based On Interestingness

主要讨论在大型数据库中挖掘多值关联规则。在对数值属性区域划分进行研究,提出多值区间下的Apriori定理、前件子集定理和后件子集定理,提出规则兴趣度的测量方法并且基于规则的兴趣度损失最小化来为区间合并作出决策。该区间合并的方法是一种全局的方法,因此可以得到更有价值的规则。

This paper does some research on Quantitative Association Rules mining technologies.After research on interval problem,an extention of Apriori Theorm is given,this paper give two theorm about the relation which have a useful guide for mering the interval and deleting the redundant rules.Lastly,present a global method based on the maximum increase of interestingness.