Mining Inconsistent Data with the Bagged MLEM2 Rule Induction Algorithm

In this paper, we investigate mining of inconsistent data using a successful approach to such data based on rough set theory. The main objective of this paper was to compare the quality of rule sets induced using two different approaches to inconsistency - lower and upper approximations - and three different approaches to ensemble voting - based on support, strength and majority - in the bagged MLEM2 algorithm for rule induction. Our main conclusion is that there is no significant difference in performance between one of the most successful techniques used in bagging, majority voting, and voting based on support (two-tailed Wilcoxon test, 5% level of significance).