jasp-stats
Feature request: brms
Hi JASP Team, Would it be possible to implement brms in JASP for Bayesian modeling within JASP? The No U-Turn Sampler (NUTS) for MCMC is extremely fast and efficient and the brms front end to the Stan language makes everything much easier for the Bayesian analyst than using Stan or using JAGS.
Thank you for recently adding JAGS. It would be great to add brms in the next JASP version. Gracias, klc108
Hi klc108,
Yes, this is very much on our agenda. We have some technical issues to sort out -- brms does not use precompiled Stan code so we'll have to figure out how to get the compiler onto people's computers. We have some bigger fish to fry first but this will definitely happen. I'll leave this issue open until we have it implemented.
Cheers, E.J.
Thank you, EJ. That's great news!
All the best, Ken (klc108)
Kenneth L. Cavanaugh, Ph.D. Research Professor Institute of Science, Technology & Public Policy Maharishi International University MR 774 1000 North 4th Street Fairfield, IA 52557 office: 641-209-5510
On Mon, Jun 15, 2020 at 2:21 PM EJWagenmakers [email protected] wrote:
Hi klc108,
Yes, this is very much on our agenda. We have some technical issues to sort out -- brms does not use precompiled Stan code so we'll have to figure out how to get the compiler onto people's computers. We have some bigger fish to fry first but this will definitely happen. I'll leave this issue open until we have it implemented.
Cheers, E.J.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/jasp-stats/jasp-issues/issues/802#issuecomment-644329041, or unsubscribe https://github.com/notifications/unsubscribe-auth/AP44DMGBHCB3UTQAVGXL3Q3RWZYC5ANCNFSM4NZYFUQQ .
Yes, clearly on our agenda. The BayesFactor package already implements a mixed model, but we want to look at this a little more closely. E.J.
@klc108 & @nahorp Is brms now implemented / used via bayesian regression, mixed models etc.? It seems so, since some other issues mentioning brms are closed: https://github.com/jasp-stats/jasp-issues/issues/308 https://github.com/jasp-stats/jasp-issues/issues/15
But I cannot directly spot it in the docs or the GUI and am not very familiar with all the bayesian lingo yet. rstanarm (and stanova) is mentioned int the docs however and it seems to be the alternative to brms, so implementing brms might not even be on the agenda anymore?
@tomtomme As far as I know, brms is not available yet. It is essentially an extremely flexible way to fit a huge variety of Bayesian regression models. @EJWagenmakers is this still on the agenda? I definitely hope so! :)
Best,
Tarandeep
