Fuzzy Bridge Regression Model Estimating via Simulation

Abstract

      The main problem when dealing with fuzzy data variables is that it cannot be formed by a model that represents the data through the method of Fuzzy Least Squares Estimator (FLSE) which gives false estimates of the invalidity of the method in the case of the existence of the problem of multicollinearity. To overcome this problem, the Fuzzy Bridge Regression Estimator (FBRE) Method was relied upon to estimate a fuzzy linear regression model by triangular fuzzy numbers. Moreover, the detection of the problem of multicollinearity in the fuzzy data can be done by using Variance Inflation Factor when the inputs variable of the model crisp, output variable, and parameters are fuzzed. The results were compared using standard mean squares error via simulated experiments and taking different sample sizes (20, 40, 80, and 160). The model's superiority was shown by achieving the least value of the mean squares error (MSE(, which indicated by the fuzzy bridge regression model.