EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Data-informed reservoir computing for efficient time-series prediction

Felix Köster iD, Dhruvit Patel iD, Alexander Wikner iD, Lina Jaurigue iD, Kathy Lüdge iD

Chaos: An Interdisciplinary Journal of Nonlinear Science · 01.07.2023 · 33 (7)

DOI: 10.1063/5.0152311 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Chaos: An Interdisciplinary Journal of Nonlinear Science, online erschienen am 05.07.2023. Verfasst von Felix Köster und 4 weiteren Autor:innen. Beteiligte Einrichtung: Institut for Theoretical Physics, Technische Universität Berlin, 10623 Berlin, Germany. Die Arbeit wurde bisher 41-mal zitiert (Crossref) und nennt 67 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Deutsche Forschungsgemeinschaft.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

We propose a new approach to dynamical system forecasting called data-informed-reservoir computing (DI-RC) that, while solely being based on data, yields increased accuracy, reduced computational cost, and mitigates tedious hyper-parameter optimization of the reservoir computer (RC). Our DI-RC approach is based on the recently proposed hybrid setup where a knowledge-based model is combined with a machine learning prediction system, but it replaces the knowledge-based component by a data-driven model discovery technique. As a result, our approach can be chosen when a suitable knowledge-based model is not available. We demonstrate our approach using a delay-based RC as the machine learning component in conjunction with sparse identification of nonlinear dynamical systems for the data-driven model component. We test the performance on two example systems: the Lorenz system and the Kuramoto–Sivashinsky system. Our results indicate that our proposed technique can yield an improvement in the time-series forecasting capabilities compared with both approaches applied individually, while remaining computationally cheap. The benefit of our proposed approach, compared with pure RC, is most pronounced when the reservoir parameters are not optimized, thereby reducing the need for hyperparameter optimization.

Autor:innen und Institutionen

  • Institut for Theoretical Physics, Technische Universität Berlin, 10623 Berlin, Germany.
  • Institute for Research in Electronics and Applied Physics, University of Maryland, College Park, Maryland 20742, USA.
  • Institute of Physics, Technische Universität Ilmenau, Weimarer Straße 25, 98693 Ilmenau, Germany.

ORCID-Profile: Felix Köster, Dhruvit Patel, Alexander Wikner, Lina Jaurigue, Kathy Lüdge

Förderung

  • Deutsche Forschungsgemeinschaft

Zitieren

APA

Köster, F., Patel, D., Wikner, A., Jaurigue, L., & Lüdge, K. (2023). Data-informed reservoir computing for efficient time-series prediction. Chaos: An Interdisciplinary Journal of Nonlinear Science, 33(7). https://doi.org/10.1063/5.0152311

Vancouver

Köster F, Patel D, Wikner A, Jaurigue L, Lüdge K. Data-informed reservoir computing for efficient time-series prediction. Chaos: An Interdisciplinary Journal of Nonlinear Science. 2023;33(7). doi:10.1063/5.0152311

BibTeX
@article{Kster2023,
  author = {Köster, Felix and Patel, Dhruvit and Wikner, Alexander and Jaurigue, Lina and Lüdge, Kathy},
  title = {Data-informed reservoir computing for efficient time-series prediction},
  journal = {Chaos: An Interdisciplinary Journal of Nonlinear Science},
  year = {2023},
  volume = {33},
  number = {7},
  doi = {10.1063/5.0152311}
}
Kostenlos starten →