Knowledge Discovery and Model Structure Selection Based on Rough Set Theory

To select a proper decision model in intelligent decision,a method to acquire the knowledge of model selection and select a decision model by utilizing the knowledge was proposed based on the rough set theory.With this method,the decision table of models with continuous attribute values is acquired by setting parameters in the models randomly.Then,the relation between objects in the decision table is obtained by setting errors.From the relation,the reduction of attributes for the decision table is conducted by applying the rough set theory to gain decision rules.Based on these decision rules,a proper model can be selected. The feasibility of the proposed method has been illustrated using an example.