ACBC-Adequate Association and Decision Rules Versus Key Generators and Rough Sets Approximations

Issue title: Rough Sets, Rule-Based Approaches, and Knowledge Representation

Guest editors: Davide Ciucci, Dominik Ślęzak and Marcin Wolski

Article type: Research Article

Authors: Kryszkiewicz, Marzena

Affiliations: Institute of Computer Science, Warsaw University of Technology, Nowowiejska 15/19, 00-665 Warsaw, Poland. mkr@ii.pw.edu.pl

Abstract: In this paper, we propose an ACBC-evaluation formula, which delivers a flexible way of formulating different kinds of criteria for association and decision rules. We prove that rules with minimal antecedents that fulfill ACBC-evaluation formulae are key generators, which are patterns of a special type. We also show that a number of types of rough set approximations of decision classes can be expressed based on ACBC-evaluation formulae. We prove that decision rules preserving respective approximations of decision classes are rules that satisfy an ACBC-evaluation formula and that antecedents of such optimal decision rules are key generators, too. A number of properties related to particular measures of association rules and key generators are derived.

DOI: 10.3233/FI-2016-1423

Journal: Fundamenta Informaticae, vol. 148, no. 1-2, pp. 65-85, 2016