An Approximate Attribute Reduction of Rough Set and Its Algorithm

In view of the deficiencies of attribute reduction in classic rough set, On condition that knowledge classification ability remains basically unchanged, this paper renders a new definition of the approximate attribute reduction of rough set and discuss its nature and algorithms. Theory proves that approximate attribute reduction is an extension of the traditional attribute reduction. Finally, a concrete example demonstrates the feasibility and effectiveness of approximate attribute reduction dealing with ambiguity and uncertainty of knowledge in information systems.