search

paul-buerkner / brms

brms R package for Bayesian generalized multivariate non-linear multilevel models using Stan
https://paul-buerkner.github.io/brms/
GNU General Public License v2.0
1.28k stars 187 forks source link

Allow skew normal to have correlated residuals #1706

Open bachlaw opened 1 week ago

bachlaw commented 1 week ago

Is there any appetite for letting the skew normal distribution become one of those allowed to model correlated residuals? Anything beyond the normal admittedly increases complexity, but the skew normal of course is the normal distribution with an added skew parameter. As such, it in theory would still be manageable and reasonably fast to sample, in addition to now being able to correlate with more common normally distributed variables as well.

paul-buerkner commented 1 week ago

Does Stan already have a multivariate skew normal distribution implemented? This would of course simplify the brms implementation a lot.

bachlaw commented 1 week ago

Definitely would help. Unfortunately it does not, which I guess is not surprising.

Sent from my iPhone

On Nov 16, 2024, at 6:15 PM, Paul-Christian Bürkner @.***> wrote:



Does Stan already have a multivariate skew normal distribution implemented? This would of course simplify the brms implementation a lot.

— Reply to this email directly, view it on GitHubhttps://github.com/paul-buerkner/brms/issues/1706#issuecomment-2480861917, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ACYOSUCUFYFUYTU4PJ3I6GD2A7N23AVCNFSM6AAAAABR5GKRRWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIOBQHA3DCOJRG4. You are receiving this because you authored the thread.Message ID: @.***>