APRIORI-SD: Adapting association rule learning to subgroup discovery

Abstract

This paper presents a subgroup discovery algorithm APRIORI-SD, developed by adapting association rule learning to subgroup discovery. This was achieved by building a classification rule learner APRIORI-C, enhanced with a novel post-processing mechanism, a new quality measure for induced rules (weighted relative accuracy) and using probabilistic classification of instances. Results of APRIORI-SD are similar to the subgroup discovery algorithm CN2-SD while experimental comparisons with CN2, RIPPER and APRIORI-C demonstrate that the subgroup discovery algorithm APRIORI-SD produces substantially smaller rule sets, where individual rules have higher coverage and significance. © Springer-Verlag Berlin Heidelberg 2003.