Innovations to Attribute Reduction of Covering Decision System Based on Conditional Information Entropy

Abstract

Traditional rough set theory is mainly used to reduce attributes and extract rules in databases in which attributes are characterised by partitions, which the covering rough set theory, a generalisation of traditional rough set theory, covers. In this article, we posit a method to reduce the attributes of covering decision systems, which are databases incarnated in the form of covers. First, we define different covering decision systems and their attributes’ reductions. Further, we describe the necessity and sufficiency for reductions. Thereafter, we construct a discernible matrix to design algorithms that compute all the reductions of covering decision systems. Finally, the above methods are illustrated using a practical example and the obtained results are contrasted with other results.