Evaluation of a classification rule mining algorithm based on secondary differences

Abstract

Rule mining is considered as one of the usable mining method in order to obtain valuable knowledge from stored data on database systems. Although many rule mining algorithms have been developed, almost current rule mining algorithms only use primary difference of a criterion to select attribute-value pairs to obtain a rule set to a given dataset. In this paper, we introduce a rule generation method based on secondary differences of two criteria for avoiding the trade-off of coverage and accuracy. Then, we performed an evaluation of the proposed algorithm by using UCI common datasets. In this case study, we compared the predictive accuracies of rule sets learned by our algorithm with that of three representative algorithms. The result shows that our rule mining algorithm can obtain not only accurate rules but also rules with the other features. © 2009 Springer Berlin Heidelberg.