Multiscale rough set model and optimization

Discretization of continuous-valued attributes is always one of the key problems in rough sets theory, a multiscale rough set model (MRSM) is developed that describes the discretization at multiple scales and analyzes the relation of classifications and certainty between scales. In view of the model´s efficiency and effectiveness. an optimal scale can be acquired with self-organization, self-study and self-adaption in a manner from coarse to fine or from fine to coarse as the human´s intelligent behaviors. Finally, a case study shows the performance of MRSM.