An Attribute Reduction Algorithm by Rough Set Based on Binary Discernibility Matrix

Attribute reduction is a very important part of rough set theory. However, the cost of reduct computation is highly influenced by the size of object set and attribute set, so enhancing the reduct computation efficiency is one of the major problems. This paper proposes a reduction algorithm capable of reducing the time cost by using the binary discernibility matrix. Based on analyzing the binary discernibility matrix, the concept of attribute significance is defined at two aspects, which is applied to lead the solutions to optimization. The validity and efficiency of the proposed algorithm is verified by comparing with other two reduction algorithms on various standard domains.