一种非对称距离下的粗糙粒结构选择方法OACHSSCDCSTPCD

Aiming at the selection of granular structures for approximate solutions in target information systems,a method for selecting rough granular structures based on asymmetric distance is proposed.Firstly,an asymmetric knowledge distance is defined to describe the closeness between conditional knowledge and the target concept within the knowledge distance framework of quotient space.Subsequently,asymmetric knowledge distance sequences are constructed according to different orders of adding conditional attributes in the same target information system.Furthermore,an evaluation parameter for rough granular structure is introduced under the constraint of attribute cost.Finally,the rough granular structure with smaller evaluation parameters is selected for step-by-step approximation.