English
 
Help Privacy Policy Disclaimer
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT
  Classification strategies in machine learning techniques predicting regime changes and durations in the Lorenz system

Brugnago, E. L., Hild, T. A., Weingaertner, D., & Beims, M. W. (2020). Classification strategies in machine learning techniques predicting regime changes and durations in the Lorenz system. Chaos, 30(5). doi:10.1063/5.0003892.

Item is

Files

show Files

Locators

show

Creators

show
hide
 Creators:
Brugnago, Eduardo L.1, Author
Hild, Tony A.1, Author
Weingaertner , Daniel1, Author
Beims, Marcus W.2, Author           
Affiliations:
1external, ou_persistent22              
2Max Planck Institute for the Physics of Complex Systems, Max Planck Society, ou_2117288              

Content

show
hide
Free keywords: -
 Abstract: In this paper, we use machine learning strategies aiming to predict chaotic time series obtained from the Lorenz system. Such strategies prove to be successful in predicting the evolution of dynamical variables over a short period of time. Transitions between the regimes and their duration can be predicted with great accuracy by means of counting and classification strategies, for which we train multi-layer perceptron ensembles. Even for the longest regimes the occurrences and duration can be predicted. We also show the use of an echo state network to generate data of the time series with an accuracy of up to a few hundreds time steps. The ability of the classification technique to predict the regime duration of more than
11 oscillations corresponds to around
10 Lyapunov times.

Details

show
hide
Language(s):
 Dates: 2020-05-042020-05-01
 Publication Status: Issued
 Pages: -
 Publishing info: -
 Table of Contents: -
 Rev. Type: -
 Identifiers: ISI: 000532279300001
DOI: 10.1063/5.0003892
 Degree: -

Event

show

Legal Case

show

Project information

show

Source 1

show
hide
Title: Chaos
  Other : Chaos : an interdisciplinary journal of nonlinear science
Source Genre: Journal
 Creator(s):
Affiliations:
Publ. Info: Woodbury, NY : American Institute of Physics
Pages: - Volume / Issue: 30 (5) Sequence Number: - Start / End Page: - Identifier: ISSN: 1054-1500
CoNE: https://pure.mpg.de/cone/journals/resource/954922836228