Identification of Non-linear Dynamic System
Ventseslav Kirilov Shopov, Vanya Dimitrova Markova · 2019
The behaviour of non-linear dynamic systems is studied. In this paper, the authors investigate the modelling and prediction abilities of a Recurrent Neural Network, Long Short Term Memory and Gated Recurrent Unit networks. The the input data sets has a chaotic nature. The effectiveness of all networks in modelling the several chaotic attractors is studied. And a comparison of their prediction quality is made.