Methods for Lower Approximation Reduction in Inconsistent Decision Table Based on Tolerance Relation

Abstract

It is well known that most of information systems are based on tolerance relation instead of the classical equivalence relation because of various factors in real-world. To acquire brief decision rules from the information systems, lower approximation reduction is needed. In this paper, the lower approximation reduction is proposed in inconsistent information systems based on tolerance relation. Moreover, the properties are discussed. Furthermore, judgment theorem and discern i bility matrix are obtained, from which an approach to lower reductions can be provided in the complicated information systems.