In covering information systems,to explore multigranularity reduction structures,this pa-per calculates pessimistic multigranularity reductions for covering information systems through dis-cernibility matrices.First,we introduce the definition of attribute reduction based on pessimistic mul-tigranulation rough sets in covering information systems.Second,we construct the discernibility ma-trices and discernibility functions.By performing disjunction and conjunction operations on the dis-cernibility matrix,all pessimistic multigranulation reductions of covering information systems can be precisely obtained.Finally,numerical experiments using 10 datasets from UCI are given to verify the feasibility and effectiveness of the proposed algorithms.
关键词
多粒度/覆盖信息系统/辨识矩阵/属性约简
Key words
multigranulation/covering information system/discernibility matrix/attribute reduction
维普
万方数据