Uncertainty Measure of Knowledge and Rough Set Based on Maximal Consistent Block Technique

In incomplete information systems, similarity measures or tolerance relations replace indiscernible relations, and the corresponding similarity or tolerance classes form coverage instead of classification of Universe. On the other hand, without satisfying the properties of transference and symmetry, there may have misjudgments in tolerance or similarity classes. Therefore, it is necessary to study roughness of knowledge and rough set based on suitable knowledge granularity in incomplete information systems. The present paper proposes a method to measure uncertain knowledge and rough set according to maximal consistent block technique, which provides the basic knowledge granulation from the similarity classes without changing the relevant model. Moreover, some new definitions about the roughness of knowledge and rough set are also discussed in the proposed method.