JumpDiff

Overview

Jump-diffusion processes are stochastic processes that include a jump term in addition to a fluctuating diffusion term.
This jump term accounts for discontinuities in the time series due to abrupt changes in the stochastic process, e.g., a sudden change in the output of a PV panel caused by the passing of clouds or in option pricing in economics.

The Python package JumpDiff uses non-parametric estimators to derive the parameters of jump-diffusion processes and thus offers methods to investigate time-series data where one has to consider jumpy behavior.
It was developed at the Institute of Energy and Climate Research at the Forschungszentrum Jülich GmbH.
The repository is available at github.com/LRydin/jumpdiff and published under an MIT-License.

A documentation with instructions for installation and usage is accessible on PyPI (jumpdiff.readthedocs.io).

A more detailed description of the package and the underlying theory is published in the peer-reviewed article in the Journal of Statistical Software (DOI:0.18637/jss.v105.i04).

Citation

If you use JumpDiff in your research, please cite it as:

Rydin Gorjão, L., Witthaut, D., and Lind, P. G. (2023). jumpdiff: A Python Library for Statistical Inference of Jump-Diffusion Processes in Observational or Experimental Data Sets. Journal of Statistical Software, 105, 1-22. DOI: 10.18637/jss.v105.i04

Acknowledgements

The authors acknowledge the funding of the Helmholtz Association via the Initiative Energy System 2050 - A Contribution of the Research Field Energy and via grant No. VH-NG-1025, and STORM - Stochastics for Time-Space Risk Models project of the Research Council of Norway (RCN) No. 274410.