Prediction of Chaos Time Series Using Elman Neural Networks
Xing Zhang · Huadong Li-Gong Daxue xuebao · 2002
This paper uses the improved Elman neural networks to predict three typical chaos time series under different noise conditions. It also discusses the relationship between learning and generalization of the neural networks and gives the optimal number of Elman's hidden layer units. In addition, the prediction results are evaluated by three targets which show the perfect performance of Elman networks in the prediction of the chaos time series.