Design of a radial basis function neural network with a radius-modification algorithm using response surface methodology

摘要: A radial basis function (RBF) neural network was designed for time series forecasting using both an adaptive learning algorithm and response surface methodology (RSM). To improve the traditional RBF network‘s capability, generalized delta rule method employed to modify radius of kernel function. Then RSM utilized explore mean square error so that appropriate combination parameters, such as number hidden nodes initial rates, could be found. Extensive studies were performed on effect values connection weights accuracy backpropagation in training artificial network. The effectiveness with proposed radius-modification technique demonstrated example intensity pulsations a laser. It found that, by utilizing techniques, provided more accurate prediction response.