This package is a Julia fork of the original author's R package (copent) and the Python package (pycopent)
The function copula_entropy
in the package indeed returns the copula entropy instead of the negative copula entropy, which is done by the corresponding copent
function in copent or pycopent.
using Distributions, CopEnt
ρ = 0.6
μ = zeros(2)
Σ = [1 ρ; ρ 1]
x = rand(MvNormal(μ, Σ), 2_000)
mi = -copent(x) # this gives us 0.1849467947096306
# true value is 0.2231435513142097
Also see the example notebook for some visualizations.