Response surface methodology for optimal neural network selection

A multilayer neural network was designed for time series forecasting using response surface methodology (RSM). To optimize the network´s parameters (the number of hidden nodes, the initial learning rate and momentum constant) RSM was employed to explore the mean square error response surface. Extensive studies were performed on the effect of the initial values of connection weights on the accuracy of the backpropagation learning method which was employed in the training of the artificial neural network. The effectiveness of the neural network with the proposed RSM technique is demonstrated with an example of forecasting the number of passengers on an international airline. It was found that with RSM the neural network provided a more accurate prediction of the response