A novel self-constructing Radial Basis Function Neural-Fuzzy System

► This paper proposes a novel self-constructing least-Wilcoxon generalized Radial Basis Function Neural-Fuzzy System (LW-GRBFNFS). ► A self-constructing scenario for generating antecedent part with PSO. Least-Wilcoxon (LW) norm is employed for estimating the accuracy of consequent part.► ► The proposed LW-GRBFNFS can provide flexible and dynamic ability for generating more accurate RBFNFS with less sensitivity to noise and outliers.