Abstract: Granular computing is a method of multilayer granular structure analysis based on problem solving,pattern classification and information processing.It is a new multidisciplinary cross discipline between rough sets,fuzzy sets,data mining and artificial intelligence.Some important properties of granular computing were discussed as well as the reduction algorithm.The traditional reduction algorithm based on granular computing is gradually calculated with reduction Core(A),but some information systems may have no reduction core.In this case,an improved reduction algorithm based on attribute significance of granular computing was proposed.The algorithm can be used in system with both reduction core and no reduction core.Finally,experiments show the feasibility of the algorithm.

Key words: Granular computing,Knowledge granule,Attribute significance,Attribute reduction