Matrix computation for dynamic modification of rough set information system

For online learning of a rule-based knowledge system, algorithms for dynamic modification of rules are needed. This paper proposes strategies of rule updating in rough set information system. Based on the concept of equivalence matrix that expresses equivalence relation, matrix algorithms for induction and modification of decision rules are given. The relations between equivalence matrix and equivalence classes are discussed. Complexity analysis shows that these algorithms are efficient for rule adjusting and unnecessary computation can be avoided.