A Delay-Based Reservoir Computing Model for Chaotic Series Prediction
Antonia Pavlidou, Xiangpeng Liang, Hadi Heidari · 2022 29th IEEE International Conference on Electronics, Circuits and Systems (ICECS) · 2022
Conventional computers based on Von Neumann architecture are unable to process complex sequential data with high efficiency. This work investigates why Delay-based Reservoir Computing (DRC) is preferred this architecture, by performing Mackey-Glass chaotic time series prediction. The outcome of the prediction and the role of the Memory Capacity (MC) for such system are presented, with simulations done in MATLAB Simulink. This developed algorithm performs with a Mean Square Error (MSE) of$\mathbf{1.2314\times 10^{-3}}$per predicted digit.