Rough Sets, Coverings and Incomplete Information

摘要: Rough sets are often induced by descriptions of objects based on the precise observations an insufficient number attributes. In this paper, we study generalizations rough to incomplete information systems, involving imprecise The role covering-based approximations that extend standard in presence about attribute values is described. setting, a covering encodes set possible partitions objects. A natural semantics two generalisations case (or non transitive tolerance relation) laid bare. It shown uncertainty due granularity description attributes and superposed, whereby upper lower themselves (in Pawlak's sense) become ill-known, each being bracketed nested sets. notion measure accuracy extended generalization construct fuzzy mappings outlined.