Knowledge Reduction in Incomplete Systems Based on γ-Tolerance Relation

摘要

The traditional rough set theory is a powerful tool to deal with complete information system, and its performance to process incomplete information system is weak, M.Kryszkiewcz has put forward the tolerance relation to handle the problem. however,the method may not be perfect on account of excessively many intersectional elements between classifications. This paper improves the tolerance relation proposed by M.Kryszkiewcz to obtain the γ — tolerance relation and γ — tolerance classes, presents rough set model for incomplete information system based on the γ — tolerance relation. The method of γ — tolerance relation is proved to be more superior to that of M.Kryszkiewcz's tolerance relation. Finally, the conception of γ — attributes reduction is defined, and the algorithm of γ — attribute reduction is provided.