Recurrent NN model for chaotic time series prediction
Jun Zhang, K.S. Tang, K.F. Man · 2002
A new Elman neural network learning algorithm is proposed for chaotic time series prediction. This method has a number of advantages over the use of a standard backpropagation algorithm. It is not only its capability for handling a much higher complexity time data series, but its superiority in time convergence can prove to be a valuable asset for time critical applications. Furthermore, this method is also very accurate in prediction as it can reach global minimum in a much attainable manner.