Hi,

Does the absence of a function `bernoulli_logit_glm_rng`

imply that something like `bernoulli_logit_rng`

should be used for generated quantities?

Side note: Version 2.28 of the Stan Function Reference appears to have duplicate definitions of the `bernoulli_logit_glm_lpmf`

and `bernoulli_logit_glm_lupmf`

functions, here:

13.3 Bernoulli-logit generalized linear model (Logistic Regression) | Stan Functions Reference

Thanks!

1 Like

Yes, you probably need to use `bernoulli_logit_rng`

- it however appears the `_rng`

version exists in the math library - math/bernoulli_logit_glm_rng.hpp at 92075708b1d1796eb82e3b284cd11e544433518e · stan-dev/math · GitHub so it is probably a simple omission that it is not exposed.

Maybe @rok_cesnovar or @stevebronder would know who to ask about this?

1 Like

As far as I know there is no reason to not add it, so just oversight. Made an issue so that we don’t repeat that mistake: Expose bernoulli_logit_glm_rng · Issue #1031 · stan-dev/stanc3 · GitHub

Thanks for digging that up Martin!

1 Like