Reductions of Information Granules in Covering Granular Computing Model

Abstract

Covering generalized rough sets model is an important extension of granular computing model based on rough sets. Based on the definition of different cover upper approximations, seven classes of coverage models are generated. The definition of original cover reduction is only suitable for partial covering model. In this paper, four kinds of cover reduction definitions are proposed, and the relation between the classes of covering reduction and the corresponding cover upper approximations is studied, and the computing method of N-reduction is given in detail. The reduction of a covering also provides a technique for data reduction in data mining.