Julia package SDE.jl¶
The main module
- Simulate diffusion processes in one or more dimension. Especially simulate vector linear processes / Ornstein-Uhlenbeck processes Monte Carlo sample diffusion bridges, diffusion processes conditioned to hit a point v at a prescribed time T Functions for transition density, mean and covariance of linear processes Perform Monte Carlo estimates of transition densities of general diffusion processes
with a Submodule
- To nonparametrically estimate the drift of a diffusion with unit diffusion coefficient using a Schauder wavelet “basis”
The package contains the additional modules:
- Alternative API to generate Ito processes and diffusions
- Random symmetric, positive definite, stable matrix for testing purposes.
The producure in
SDE.Schauder is a Julia implementation of nonparametric Bayesian inference for
“continuously” observed one dimensional diffusion processes with unit diffusion coefficient. The drift
is modeled as linear combination of hierarchical Faber–Schauder basis functions with a Gaussian prior
on the coefficients. This incorporates a Brownian motion like prior on the drift function. The posterior is
then computed using Gaussian conjugacy.
This is work in progress.