The Gibbs sampler discussed on Darren Wilkinson's blog and also on Dirk Eddelbuettel's blog has been implemented in several languages, the first of which was R.
The task is to create a Gibbs sampler for the unscaled density
f(x,y) = x x^2 \exp(-xy^2 - y^2 + 2y - 4x)
using the conditional distributions
x|y \sim Gamma(3, y^2 +4)
y|x \sim Normal(\frac{1}{1+x}, \frac{1}{2(1+x)})
| using CSV, DataFrames, Downloads, Tar | |
| datadir = "biofast-data-v1" | |
| tarball = "$datadir.tar.gz" | |
| if !isfile(tarball) | |
| dataurl = joinpath( | |
| "https://github.com/lh3/biofast/releases/download", | |
| datadir, | |
| tarball, | |
| ) |
| --- | |
| jupyter: julia-1.7 | |
| --- | |
| ```{julia} | |
| using RCall | |
| ``` | |
| ```{julia} | |
| RCall.ijulia_setdevice(MIME("image/svg+xml")) |
These notes are based on Patrick O'Leary's video that he kindly created when I was wingeing about problems with git. It is best to watch the video first then go back through these notes to recall the steps.
Create a clone of the JuliaLang repository
git clone git://github.com/JuliaLang/julia.gitor
git clone git://github.com/JuliaLang/julia.git myjulia