Attribute Reduction Based on Optimistic Multi-granulation Information Systems

Attribute reduction is the most important and widely applied part in rough sets. Multi-granulation rough set model is a significant generalization of classical rough sets, which includes pessimistic and optimistic multi-granulation models. Several attribute reduction algorithms based on multi-granulation models are designed in literatures, but all of them are based on pessimistic models, while the attribute reduction based on optimistic models has not been developed. Thus, in this paper, we propose an attribute reduction approach, named related family, for the first and the second optimistic multi-granulation covering rough set models, which is the basis of attribute reduction of all optimistic multi-granulation rough set models and decrease the time complexity of attribute reduction.