Detecting metastable states of dynamical systems by recurrence-based symbolic dynamics

Abstract : We propose an algorithm for the detection of recurrence domains of complex dynamical sys- tems from time series. Our approach exploits the characteristic checkerboard texture of recurrence domains exhibited in recurrence plots (RP). In phase space, RPs yield intersecting balls around sampling points that could be merged into cells of a phase space partition. We construct this parti- tion by a rewriting grammar applied to the symbolic dynamics of time indices. A maximum entropy principle defines the optimal size of intersecting balls. The final application to high-dimensional brain signals yields an optimal symbolic recurrence plot revealing functional components of the signal.
Document type :
Journal articles
Physical Revew Letters, APS, 2013, 110, pp.154101. <10.1103/PhysRevLett.110.154101>


https://hal.inria.fr/hal-00847164
Contributor : Axel Hutt <>
Submitted on : Monday, July 22, 2013 - 6:17:42 PM
Last modification on : Tuesday, September 3, 2013 - 10:41:02 AM

File

RecuParti.pdf
fileSource_public_author

Identifiers

Collections

Citation

Peter Beim Graben, Axel Hutt. Detecting metastable states of dynamical systems by recurrence-based symbolic dynamics. Physical Revew Letters, APS, 2013, 110, pp.154101. <10.1103/PhysRevLett.110.154101>. <hal-00847164>

Export

Share

Metrics

Consultation de
la notice

214

Téléchargement du document

73