Abstract:

Multi-granularity is an effective data processing method in rough set theory. Granularity reduction is the prerequisite for obtaining the concise rules of the information system. The granulation reduction of graded optimism (pessimism) multigranulation rough set is researched and the lower approximation definition of graded rough set is improved. The granulation matrix of graded multigranulation rough set is proposed. Based on the granulation matrix, the lower approximation calculation and necessity of granularity are studied in graded multigranulation rough set. Then the core granulation definition on graded optimism multigranulation rough set is given. The granulation reduction of graded optimism (pessimism) multigranulation rough set is proposed based on granulation matrix. Finally, a numerical example is given to demonstrate the correctness of the proposed method for granularity reduction.