search

stan-dev / rstanarm

rstanarm R package for Bayesian applied regression modeling
https://mc-stan.org/rstanarm
GNU General Public License v3.0
389 stars 134 forks source link

default degrees of freedom for hs() and hs_plus() #369

Closed mcol closed 5 years ago

mcol commented 5 years ago

Summary:

The documentation asserts that default degrees of freedom for hs and hs_plus are 3, but in the code they are actually set to 1.

RStanARM Version:

git repository

mcol commented 5 years ago

I can submit a patch for this, but I'm not sure if I should fix the code or the documentation. I think 3 degrees of freedom was the suggested value for the non-regularized horseshoe, but perhaps it's still a safe default even for the regularized horseshoe?

avehtari commented 5 years ago

Thanks for pointing this out (and coming back to remind us!) Please, fix the documentation. 1 is the correct one, 3 was something we experimented to reduce divergences, but then regularized hs with additional slab component turned out to be a better solution.

jgabry commented 5 years ago

Closing this since I just merged #373. Thanks for catching this and fixing the doc!