Research on Positive and Negative Association Rules Based on "Interest-Support-Confidence" Framework
Yanguang Shen, Jie Liu, Zhiyong Yang · 2009
The traditional method based on "support-confidence" framework could bring a large number of irrelevant or even misleading association rules. In view of the problem of evaluation standards of association rules, we increase interest measure in the evaluation standards, and give the definition of interest measure and positive and negative association rules method based on "interest-support-confidence" framework, which can be used to mine negative association rules. We applied this method in data mining of e-business , and through comparing with the Apriori method, this method can effectively reduce the amount of positive association rules, and produce more meaningful negative association rules.