Universal approximation theorem for a radial basis function fuzzy neural network

Radial basis function fuzzy neural network has one hidden layer. Inputs are used to find the mean and variance of the Gaussian activation function. In this article, the hidden layer weights which are taken as fuzzy numbers and the activation function values are used to determine the output function. The method of finding the output of this network is illustrated by an example. The output function of radial basis function fuzzy neural network act as a function approximator. Further, Universal approximation theorem for radial basis function fuzzy neural network is dealt with. This is proved using Stone Weierstrass Theorem. Copyright (C) 2022 Elsevier Ltd. All rights reserved.