Biqing Wang: Quick Attribute Reduction Algorithm Based on Consistent Degree

As is well known, attribute reduction is one of the basic concepts in rough set theory. In this paper, a new algorithm for computing equivalence classes based on segment quick sort is provided and its time complexity is cut down to O(|C||U|) compared with traditional algorithms. Then a method for computing object number of positive region without computing positive region is given. On this basis, the concept of consistent degree is introduced and its properties are studied. Furthermore, definitions of core attributes and attribute importance are proposed. The algorithm for computing core attributes and attribute importance are designed. Finally, a quick attribute reduction algorithm which uses core attributes as initial reduction and attribute importance as heuristic information is presented. In the worst case, its time complexity is O(|C|2|U|). Example analyses and experimental results show that the quick attribute reduction algorithm of this paper is feasible and efficient.