Knowledge Reduction of Covering Approximation Space

Covering approximation space is a kind of knowledge representation different from Pawlak´s approximation space, and knowledge reduction is the key step in knowledge acquisition. Zhu proposed an absolute reduction of covering approximation space, but it could only reduce absolutely redundant knowledge. In order to reduce relatively redundant knowledge with respect to a decision, the problem of relative reduction is studied in this paper. We find that the rough approximations keep unchanged in the reduced space. In addition, an algorithm for knowledge reduction of covering approximation space is proposed. It can reduce not only absolutely redundant knowledge but also relatively redundant knowledge.