A Core-Reduction Algorithm Based on an Improved Discernible Table

Knowledge reduction based on rough sets theory is an important but NP-hard problem.Among the present various algorithms of knowledge reduction based on rough sets model,core is always initialized directly in knowledge reduction as attribute reduction set,which can efficiently reduce the searching scope of reduction algorithm in attributes space and accelerate the carrying out of knowledge reduction to some degree.However,the conclusion of core attribute is basically obtained by using Hu discernible matrix.The present paper,considering Hu's discernible matrix algorithm,is to discuss the problems existing in Hu and the algorithm of Wroblewaski discernible table,better the definition of discernible table and together with Ye method,and puts forward a core-reduction algorithm based on discernible table.In this way,it not only overcomes the disadvantage of methods of discernible matrix,but also avoids the problems in algorithm of discernible table.Result of experiment proves that this algorithm can effectively work out the attribute core in decision-making system.