文摘
Multigranulation rough set is a novel generalization of Pawlak’s rough set through using multiple granular structures instead of single granular structure. By considering the maximal and minimal operators used in the optimistic and pessimistic multigranulation fuzzy rough sets, we devote to present an adjustable multigranulation fuzzy rough set. Such new model is constructed on a parameterized binary operator, which is an improvement of maximal and minimal operators. It is shown that both optimistic and pessimistic multigranulation fuzzy rough sets are special cases of adjustable multigranulation fuzzy rough set. Moreover, we also derive an approximation quality significance measure and design a forward greedy algorithm for granular structures selection. Experiments show the validity of the proposed algorithm from search strategy in the meaning of parameters used in adjustable multigranulation fuzzy rough sets.