Time series prediction with simple recurrent neural networks

Abstract

Simple recurrent neural networks are widely used in time series prediction. Most researchers and application developers often choose arbitrarily between Elman or Jordan simple recurrent neural networks for their applications. A hybrid of the two called Elman-Jordan (or Multi-recurrent) neural network is also being used. In this study, we evaluated the performance of these neural networks on three established bench mark time series prediction problems. Results from the experiments showed that Jordan neural network performed significantly better than the others. However, the results indicated satisfactory forecasting performance by the other two neural networks.

Key Words: Time Series Prediction, Artificial Neural Network, Recurrent NN, Resilient Propagation.