Semi-monolayer Cover Rough Set: Concept, Property and Granular Algorithm.

Zhengjiang Wu,Ning Chen,Yan Gao

DOI: https://doi.org/10.1016/j.ins.2018.04.066

IF: 8.1

2018-01-01

Information Sciences

Abstract:Rough set theory is known as a powerful tool for dealing with incomplete, inconsistent and insufficient information in decision making. With the constraint that each set of cover has at least one reliable element, the semi-monolayer cover (SMC) is a special neighbor-hood system defined on the universe between the partition and regular cover. More similar to Pawlak approximation operators than cover approximation operators, SMC approximation operators have some interesting properties, such as the equivalent granular forms, the cross duality, etc. In this paper, we define four pairs of SMC rough approximation operators, namely GC, GU, DE and DA, and then discuss their properties and relations. It has been proven that (1) GU is the equivalent granular-form approximation operator of DE. (2) DE and DA are two pairs of cross dual approximation operators. That is to say that upper DE and lower DA, lower DE and upper DA are two pairs of dual approximation operators. The equivalent relation and cross dual relation serve as a bridge connecting granular-form approximation operators with the dot-form ones. (3) A DE or DA approximation set can be rebuilt by the corresponding GC approximation set and an SMC-Mapping. The equivalent granular form of DE and DA gives rise to some advanced algorithms that improve the efficiency of calculating DE and DA approximation sets. Experimental results on a UCI data set show the granular algorithm achieves better performance than the basic dot algorithm, and the speedup of efficiency tends to be exponential with the size of universe. (C) 2018 Elsevier Inc. All rights reserved.