General fuzzy regression using least squares method

This article introduces a general fuzzy regression model, which separates the response function on a mode and spreads of an alpha-level set for an observed fuzzy number, to estimate a fuzzy relation between two fuzzy random variables. We construct the general fuzzy regression model using least squares estimation and best response functions on the mode and spread of an alpha-level set for the fuzzy number when the response variable is an LR-fuzzy number and independent variables are crisp numbers. Then we derive a crisp mean and variance of the predicted fuzzy number, and compare the accuracy of our proposed fuzzy regression model with other fuzzy regression models suggested by many authors.