kramersmoyal: Kramers--Moyal coefficients for stochastic processes

TL;DR

kramersmoyal is a Python library that uses kernel-density estimators to extract Kramers--Moyal coefficients, including drift and diffusion, from stochastic time series of any dimension and order.

Contribution

It introduces a non-parametric method for estimating Kramers--Moyal coefficients from multivariate stochastic data using kernel density estimation.

Findings

Enables extraction of drift and diffusion coefficients from complex time series.

Supports any dimension and order of stochastic processes.

Provides a flexible, non-parametric estimation approach.

Abstract

kramersmoyal is a python library to extract the Kramers--Moyal coefficients from timeseries of any dimension and to any desired order. This package employs a non-parametric Nadaraya--Watson estimator, i.e., kernel-density estimators, to retrieve the drift, diffusion, and higher-order moments of stochastic timeseries of any dimension.