An algorithm for reusable uninteresting rules in association rule mining

Abstract

In this paper, we present a new framework for reusable association rule mining based on chi2 and odds ratio. We start at mining the association rules using standard Apriori algorithm. The strong rules are defined as association rules, while the weak rules will be evaluated by our proposed method. Firstly, the weak rules must be converted to 2 times 2 contingency table. We then compute the relationship between variables using chi2 and odds ratio. If the weak rules are related to each other with positive or negative relationship, then the weak rules will also be determined as association rules. Our system is evaluated with experiments on the crime data.