NoLiTSA (NonLinear Time Series Analysis) is a Python module implementing several standard algorithms used in nonlinear time series analysis.

• Estimation of embedding delay using autocorrelation, delayed mutual information, and reconstruction expansion.
• Embedding dimension estimation using false nearest neighbors and averaged false neighbors.
• Computation of correlation sum and correlation dimension from both scalar and vector time series.
• Estimation of the maximal Lyapunov exponent from both scalar and vector time series.
• Generation of FT, AAFT, and IAAFT surrogates from a scalar time series.
• Simple noise reduction scheme for filtering deterministic time series.
• Miscellaneous functions for end point correction, stationarity check, fast near neighbor search, etc.

NoLiTSA can be installed using pip by running

git clone https://github.com/manu-mannattil/nolitsa.git
pip install ./nolitsa

or, with newer versions of pip:

pip install git+https://github.com/manu-mannattil/nolitsa.git

NoLiTSA requires NumPy and SciPy, and (theoretically) works with both Python 2 and 3.

NoLiTSA’s unit tests can be executed by running nosetests, or by using

python setup.py test

Versions of NoLiTSA were used in the following publications:

• M. Mannattil, H. Gupta, and S. Chakraborty, “Revisiting Evidence of Chaos in X-ray Light Curves: The Case of GRS 1915+105,” Astrophys. J. 833, 208 (2016).

• M. Mannattil, A. Pandey, M. K. Verma, and S. Chakraborty, “On the applicability of low-dimensional models for convective flow reversals at extreme Prandtl numbers,” Eur. Phys. J. B 90, 259 (2017)

Sagar Chakraborty is thanked for several critical discussions.

NoLiTSA is licensed under the 3-clause BSD license. See the file LICENSE for more details.