Dynamic System Modeling Based on Recurrent Neural Network

Recurrent neural networks are widely used in time series prediction and classification. However, it has problems such as insufficient memory ability and difficulty in gradient back propagation. To overcome this drawback, this article proposed a new recurrent neural network model called RNN-SKIP. It can strengthen the ability to remember mformatlon from past moments and help the gradient to propagate backwards more smoothly. By testing arrhythmia data and analyzing the model effects of different parameters through experiments, we found that the new RNN-SKIP model can optimize the structure and improve the accuracy of the recurrent neural network, and effectively solve the exploding gradient and vanishing gradient problem.