Question regarding the support of the Geometric distribution in Julia

General Usage
1

Hi all,

I am evaluating a likelihood involving the Geometric distribution, with a binary latent variable which indicates whether to set the parameter value p of the Geometric distribution to be equal to p=1.

As these terms when p=1 always evaluate to 1, it is not a strong requirement that I have to use something like logpdf to evaluate these observations, but I don’t understand why the support for the Geometric distribution within the Distributions package is 0<p<1 opposed to 0<p\leq1.

Thanks in advance.

2

Does it work if you set check_args=false?

I briefly looked at the source code and saw this:

https://github.com/JuliaStats/Distributions.jl/blob/a6b99600a4d3182e533bb3bb4e0358f6aff50c8d/src/univariate/discrete/geometric.jl#L32

But haven’t tested.