frappy (fractal and phase space analysis for python) is a Python package implementing a non subjective algorithm for calculating correlation dimensions from data.
This package currently contains functions for:
- Converting a time series into a uniform deviate through a rank transformation
- Embedding time series into a d-dimensional vector space
- Calculate correlation dimension at different embedding dimensions
- Calculate saturated correlation dimension and minimum embedding dimension using curve fitting
- Module for generating surrogate data using
- Random shuffling
- Preserving autocorrelation at lag-1 (AR(1) surrogates)
You can install frappy from PyPi using
pip install frappy
This work is an python implementation of the algorithm described in Harikrishnan, K. P., Misra, R., Ambika, G., & Kembhavi, A. K. (2006). A non-subjective approach to the GP algorithm for analysing noisy time series. Physica D: Nonlinear Phenomena, 215(2), 137-145.