Attribute Reduction Based on Approximation Set of Rough Set ⋆

Attribute reduction is one of the core issues of the rough set theory. The traditional method of attribute reduction was designed based on positive region unchanged, as abandoning processing the incompatible data in the boundary domain, so the classification characteristics of the reduction results are not necessarily best. In this paper, the concept of relative approximate degree is proposed in the decisionmaking information systems on the basis of the research on approximation set of rough set, the degree of similarity between the division of domain made by condition and division attribute is well reflected. Then the heuristic reduction attribute search criteria are using the changing of the relative approximate degree, the attribute reduction algorithm based on approximation set of rough set is proposed. The experimental results show that the reduction attribute obtained by the proposed method has better classification characteristics in the various classification algorithms.