jumpdiff: A Python Library for Statistical Inference of Jump-Diffusion Processes in Observational or Experimental Data Sets
Peer reviewed, Journal article
Published version
Permanent lenke
https://hdl.handle.net/11250/3059694Utgivelsesdato
2023Metadata
Vis full innførselSamlinger
Originalversjon
Journal of Statistical Software. 2023, 105 (1), 1-22. https://doi.org/10.18637/jss.v105.i04Sammendrag
We introduce a Python library, called jumpdiff, which includes all necessary functions to assess jump-diffusion processes. This library includes functions which compute a set of non-parametric estimators of all contributions composing a jump-diffusion process, namely the drift, the diffusion, and the stochastic jump strengths. Having a set of measurements from a jump-diffusion process, jumpdiff is able to retrieve the evolution equation producing data series statistically equivalent to the series of measurements. The back-end calculations are based on second-order corrections of the conditional moments expressed from the series of Kramers-Moyal coefficients. Additionally, the library is also able to test if stochastic jump contributions are present in the dynamics underlying a set of measurements. Finally, we introduce a simple iterative method for deriving secondorder corrections of any Kramers-Moyal coefficient.