Abstract: Some contradictive rules would be often mined when the traditional support and confidence mode is adopted to do the positive and negative association rules mining. This paper presents the concept of upgrade rate and elaborates its significance and function. The mode based on support and upgrade rate for positive and negative association rules mining is also provided. What revealed by the present experiment includes that the rules mined by adopting this mode in the positive and negative association rules mining have the advantages of less quantity, higher quality and no false rules and contradictive rules.

摘要： 在进行正负关联规则挖掘时，传统的支持度—置信度模式经常会挖掘出相互矛盾的规则。本文提出了提升率的概念，并阐述了提升率的意义和作用。提出了基于支持度—提升率的正负关联规则挖掘模式。实验表明，采用该模式进行正负关联规则挖掘，所挖掘出的规则数量少，质量高，无虚假规则和相互矛盾的规则。