ogunjosam / chaos-analysis

A repository with several methods for analysis of experimental time series of different nature.

chaos-analysis

A code implementing several methods for analysis of chaotic behavior of experimental time series of different nature.

Requirements

Supports Python 3, required libraries in the file -> requirements.txt

Methods

• Fast Fourier transform
• Autocorrelation
• Embedding time
• Phase space reconstruction
• Correlation sum and correlation dimension
• Lyapunov exponents
• Kaplan–Yorke conjecture
• 0-1 Test for chaos
• Time series generation for different systems
• Classification with classical models
  • k-NN
  • Random Forest
  • Nearest Centroid
  • Naive Bayes
  • Linear SVM
• Classification with multilayer perceptron
• Classification with shallow architecture
• Classification with deep architecture

References:

1. Kantz, Holger, and Thomas Schreiber. Nonlinear time series analysis. Vol. 7. Cambridge university press, 2004.
2. Aoki, Kazunori. Nonlinear dynamics and chaos in semiconductors. CRC Press, 2000.
3. J. Awrejcewicz et. al., Entropy, 20, 175 (2018).
4. M. Mannattil et. al., Astrophys. J. 833, 208 (2016) https://github.com/manu-mannattil/nolitsa.
5. Schölzel, Christopher, Nonlinear measures for dynamical systems, Zenodo (2019) https://pypi.org/project/nolds/.
6. D. Toker et. al., Commun. Biol. 3, 11 (2020).
7. H. I. Fawaz et. al., Data Min. Knowl. Disc. 33, 917–963 (2019) https://github.com/hfawaz/dl-4-tsc.
8. G. A. Gottwald et. al., arXiv:0906.1418 (2009) http://arxiv.org/pdf/0906.1418v1.