Berlin 2024 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
DY: Fachverband Dynamik und Statistische Physik
DY 42: Focus Session: Computing with Dynamical Systems: New Perspectives on Reservoirs and Applications I – Fundamentals
DY 42.4: Talk
Thursday, March 21, 2024, 10:30–10:45, BH-N 243
Enhancing reservoir predictions of chaotic time series by incorporating delayed values of input and reservoir variables — •Luk Fleddermann1,2 and Ulrich Parlitz1,2 — 1Max Planck Institute for Dynamics and Self-Organization, Göttingen, Germany — 2Institute for the Dynamics of Complex Systems, University of Göttingen, Germany
Chaotic time series can be predicted using linear readouts of driven reservoir dynamics. In the applied case, typically only time series of incomplete measurements, i.e. partial observations of the dynamical state, are available. We compare the performance of reservoir computing for time series predictions with complete and partial system state knowledge. By combining delayed values of input and reservoir variables, we increase the mean time length of valid predictions for partial observations of the dynamical state.
Keywords: Reservoir Computing; Time Series Prediction; Delayed Values; Echo State Networks