The Utilities of Imprecise Rules and Redundant Rules for Classifiers

Abstract

Rules inferring the memberships to single decision classes have been induced in rough set approaches and used to build a classifier system. Rules inferring the memberships to unions of multiple decision classes can be also induced in the same manner. In this paper, we show the classifier system with rules inferring the memberships to unions of multiple decision classes has an advantage in the accuracy of classification.We look into the reason of this advantage from the view point of robustness. The robustness in this paper implies the maintenance of the classification accuracy against missing values. Moreover, we examine the relationship between the accuracy and the robustness via numerical experiments.We demonstrate an advantage of the redundant rules. Finally, a stronger advantage of rules inferring the memberships to unions of multiple decision classes is reexamined.