A nonlinear hybrid fuzzy least-squares regression model

A method for quadratic hybrid fuzzy least-squares regression is developed in this paper. Input and output information is presented in the form of trapezoidal fuzzy numbers. The method of regressions creation is based on the transformation of the input and output fuzzy numbers into intervals, which are called weighted intervals. The proposed method extends a group of initial data membership functions as it can be applied not only to normalized triangular fuzzy numbers, but also to trapezoidal fuzzy numbers. The numerical example has demonstrated that the developed regression model can be used for analysis of relations among qualitative characteristics and for prediction its meanings with success.