Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Reconstruction of chaotic dynamics by on-line EM algorithm.

S Ishii1, M A Sato

  • 1Nara Institute of Science and Technology, Ikoma-shi, Japan. ishii@is.aist-nara.ac.jp

Neural Networks : the Official Journal of the International Neural Network Society
|November 23, 2001
PubMed
Summary

This study introduces a normalized Gaussian network (NGnet) to reconstruct chaotic dynamics, demonstrating its robustness against system and observation noise for accurate attractor reproduction and prediction.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

The variability of baroreflex sensitivity in juvenile, spontaneously hypertensive rats.

Cardiovascular journal of Africa·2011
Same author

Changes in sodium appetite evoked by lesions of the commissural nucleus of the tractus solitarius.

Brazilian journal of medical and biological research = Revista brasileira de pesquisas medicas e biologicas·2009
Same author

Role of the medulla oblongata in hypertension.

Hypertension (Dallas, Tex. : 1979)·2001
Same author

Enhanced pressor response to carotid occlusion in commNTS-lesioned rats: possible efferent mechanisms.

American journal of physiology. Regulatory, integrative and comparative physiology·2000
Same author

Commissural NTS lesions and cardiovascular responses in aortic baroreceptor-denervated rats.

Hypertension (Dallas, Tex. : 1979)·1999
Same author

Central moxonidine on water and NaCl intake.

Brain research bulletin·1999

Area of Science:

  • Dynamical Systems and Chaos Theory
  • Machine Learning
  • Nonlinear Dynamics

Background:

  • Chaotic dynamics are complex and sensitive to initial conditions.
  • Reconstructing and predicting chaotic systems is challenging, especially with noise.
  • Existing methods may struggle with noise and partial observations.

Purpose of the Study:

  • To develop a robust method for reconstructing chaotic dynamics using a normalized Gaussian network (NGnet).
  • To evaluate the NGnet's performance under system and observation noise.
  • To assess the NGnet's ability to handle partially observed chaotic systems via delay embedding.

Main Methods:

  • Utilized a normalized Gaussian network (NGnet), a network of local linear regression units.
  • Trained the NGnet using an on-line Expectation-Maximization (EM) algorithm to learn the system's vector field.

Related Experiment Videos

  • Applied the delay embedding method for reconstructing dynamics from partially observed variables.

    • Main Results:

    • The trained NGnet successfully reproduced chaotic attractors, capturing complexity and instability.
    • The method demonstrated robustness against both system noise and observation noise.
    • Effective learning of chaotic dynamics in delay coordinate space was achieved even with noise.

    Conclusions:

    • The NGnet provides a robust approach for reconstructing chaotic dynamics.
    • The method is effective in the presence of various noise types and partial observations.
    • NGnets offer a promising tool for analyzing and predicting complex chaotic systems.