Attribute Reduction Algorithm Based on Incomplete Decision Table

Abstract

The paper describes the basic concepts of rough set theory and discernibility matrix and presents an attribute reduction algorithm based on reduced discernibility matrix, which aims at resolving the inadequate of the existing attribute reduction based on incomplete decision table. There only contain useful elements for the algorithm in the reduced discernibility matrix, which obtain one reduction of incomplete decision table by iteration and set operations. The experimental results show that the algorithm can not only obtain reduced attribute, but also reduce the computation time and storage space greatly. Keywords-incomplete decision table; attribute reduction; rough set; discernibility matrix; algorithm complexity