Memristor-based Echo State Network and Prediction for Time Series

Abstract

This paper proposes echo state network based on the memristor and applies it to the prediction of chaotic time series. The memristor is used to replace the connection synapses between neurons in the echo state network reservoir, which improves the prediction performance of the network. A simulation experiment is carried out on the echo state network based on the memristor in the widely used chaotic time series prediction. The experiment results show that the echo state network based on the memristor improves the prediction accuracy and performance of the network.