In this paper, we incorporate the concept of fuzzy set theory into the support vector regression (SVR). In our proposed method, target outputs of training samples are considered to be fuzzy numbers and then, membership function of actual output (objective hyperplane in high dimensional feature space) is obtained. Two main properties of our proposed method are: (1) membership function of actual output can be obtained without pre-assumption on type of membership function of the bias term and the components of weight vector; (2) the membership function of target output can be each type of fuzzy number.

Keywords