Hypersurface Discretization Method of Continuous Attribute in Complex Decision-Making Problems

This paper presents a hypersurface discretization method of continuous-valued attribute in complex decision-making problems. Firstly, the definition of hypersurface in decision system is given and the hypersurface is proved to be a high-order polynomial. Item sum of the polynomial is given. Meanwhile, the relation between the number of hypersurface and the maximal number of decision rules is also proved. Then a method for obtaining hypersurface by support vector machine (SVM) is proposed. Finally, taking the store strategy of low consumption items as an example, the paper compares hypersurface discretization method with hyperplane discretization method. The experiment result indicates that the indirect hypersurface discretization method is more effective to classify the decision classes in decision tables and obtain simpler decision rules.