lsp-sim 0.1.1

Locally stationary process simulation

lsp-sim

Locally Stationary Process Simulation in Python

This package provides tools to simulate locally stationary processes (LSPs) in Silverman’s sense with a flexible covariance structure: stationary correlation part $r(\tau)$ and time-varying power part $q(\eta)$, both chosen as Gaussian functions.

It includes the functions:

-lsp_f0_sim for simulating LSP realizations

-wv_lsp for calculating the Wigner-Ville distribution of the LSP

Jupyter Demo

Use the simple Jupyter notebook demo demo_lsp_sim.ipynb for exploring parameter effects.
Reproducible simulations can be obtained by fixing random seeds.
Full source code and notebook demo are available on GitHub https://github.com/RacheleAnderson/lsp-sim/

The demo visualizes:

• Simulated realizations showing different behaviors of the realizations when changing the parameters (3 cases)
• Covariance matrices for the 3 cases of different parameter configurations
• Wigner-Ville distribution for the 3 cases of different parameter configurations

Background

The model is presented in the research paper

Anderson, R., Sandsten, M. Time-frequency feature extraction for classification of episodic memory. EURASIP J. Adv. Signal Process. 2020, 19 (2020).

available online (Open Access) at: https://doi.org/10.1186/s13634-020-00681-8

Previous Matlab code is available for repeating the study, see https://github.com/RacheleAnderson/lsp-time-frequency.

This Python package is a simpler version with the limited scope of simulating LSP realizations according to the model.

License

MIT License — see LICENSE for details.