Today in maths is weird: Imagine you want to simulate vote shares for N parties. You call runif(N), then normalise. Seems reasonable, right? Nope! This is biased towards cases where vote shares are similar.
Instead, you should use the symmetric Dirichlet distribution brms::rdirichlet(10, rep(1, N))
Instead, you should use the symmetric Dirichlet distribution brms::rdirichlet(10, rep(1, N))
November 2, 2025 at 11:45 AM
Everybody can reply
7 likes
Just fit a Bayesian multilevel moderated mediation model with #brms like it was nothing. 🥳
October 28, 2025 at 10:52 AM
Everybody can reply
17 likes
#HappyBirthday Àlex Soler-Roig, 94. The son of a rich Catalan surgeon, he started 6 #F1 GPs from ’71 to ’72, in Lotuses, Marches & BRMs, but DNF’d in all 6. Pic: in a Porsche 908 at Circuito de Guadalope Alcañiz, a superfast & now disused street circuit in Aragón, Spain, in ’69.
October 29, 2025 at 7:35 AM
Everybody can reply
2 reposts
43 likes
前回のフェス絵と対比っぽくしたかったんだけど全然無理だったねの吊るし(?)
bsky.app/profile/brms...
bsky.app/profile/brms...
October 25, 2025 at 3:42 PM
Everybody can reply
7 likes
I've done some more work on the relative measurement uncertainty, comparing `brms` posterior draws to "plausible values" in Rasch models, and some other reliability metrics. Estimating RMU from draws adds some variation, as shown in the figure.
pgmj.github.io/reliability....
#rstats #psychometrics
pgmj.github.io/reliability....
#rstats #psychometrics
October 24, 2025 at 4:49 PM
Everybody can reply
4 likes
This is a good post! I keep using 1234 or 12345, but only for visualization stuff like jittering. For real stuff, I go to random dot org (based on atmospheric noise) and create a random 8+ digit integer, generally one for each brms model I run in a project
We need to have a conversation about random seeds. Don't use 42.
blog.genesmindsmachines.com/p/if-your-ra...
blog.genesmindsmachines.com/p/if-your-ra...
If your random seed is 42 I will come to your office and set your computer on fire🔥
Figuratively. More likely you'll get a stern talking to.
blog.genesmindsmachines.com
October 22, 2025 at 2:22 PM
Everybody can reply
4 reposts
2 quotes
24 likes
Project idea: Find published open data sets that do that, re-run models with appropriate random effect structures in brms, and compare how often the published claims hold. Anyone in?
Pretty wild to me how many people when realizing their linear mixed effects models do not converge with appropriate random effect structures, reduce the random effect structure until convergence and then very confidently use the inferential results to make far-reaching claims about nature.
October 22, 2025 at 7:23 AM
Everybody can reply
1 reposts
12 likes
Brms nerds!
Does anyone know a workaround for modeling residual correlations in multivariate models in {brms}? It only works for Gaussian and Student distributions right now, but maybe someone has a hack…
#RStats (Maybe @solomonkurz.bsky.social?)
#RStats (Maybe @solomonkurz.bsky.social?)
October 21, 2025 at 12:14 PM
Everybody can reply
3 likes
Andrew Johnson has worked up a copula function for brms (which is on the list for version 3 release). Here’s a link to the feature request:
github.com/paul-buerkne...
github.com/paul-buerkne...
Gaussian Copulas · Issue #1317 · paul-buerkner/brms
Hey Paul, I'm currently working on adding gaussian copulas to brms using @spinkney's implementations over in https://github.com/spinkney/helpful_stan_functions/blob/main/functions/copula/centered_g...
github.com
October 21, 2025 at 11:50 AM
Everybody can reply
3 likes
Yup! Wasn't sure if someone had already done this and made it usable via brms
bsky.app/profile/step...
bsky.app/profile/step...
That's all I can think of, but wasn't sure if someone had a workaround and/or had already done it
October 21, 2025 at 12:32 PM
Everybody can reply
1 likes
October 15, 2025 at 6:52 PM
Everybody can reply
1 likes
意思伯爵（@brms-k6-k.bsky.social ）様宅「木原工房」にお邪魔しました✨️
また、遊太さん（と大学芋）を少しだけお借りしました。
ありがとうございました🙏
店内の様子や関係性など、解釈違いありましたらすみません💦
また、遊太さん（と大学芋）を少しだけお借りしました。
ありがとうございました🙏
店内の様子や関係性など、解釈違いありましたらすみません💦
October 15, 2025 at 8:17 AM
Everybody can reply
4 likes
Started the day with this incredible 1970 footage of Can playing an 18 minute version of Oh Yeah
www.youtube.com/watch?v=bRMS...
www.youtube.com/watch?v=bRMS...
Can - Oh Yeah
YouTube video by ark80
www.youtube.com
October 11, 2025 at 5:41 AM
Everybody can reply
1 reposts
21 likes
1 saves
October 10, 2025 at 9:00 AM
Everybody can reply
11 likes
The answer with many of these things is "wait for someone to write an abstraction layer", like the amazing @paulbuerkner.com and brms team
October 4, 2025 at 8:10 PM
Everybody can reply
2 likes
Mechanics stayed up all night rebuilding Clark’s car, but the BRMs of Hill and Jackie Stewart both failed before the halfway mark.
October 2, 2025 at 5:57 AM
Everybody can reply
3 likes
September 29, 2025 at 3:57 PM
Everybody can reply
1 likes
“Everything is dirty,” he said. He sd he cleaned his cell using hs own soap, shampoo & towel, & at 1 pt detainees wer givn 2 buckets of wtr, 2 brms & mops & tld to clean a dorm housing roughly 90 ppl.
“Wld u let yr own fmly live in thes conds?”... “It’s unethical. It’s irresponsible. It’s unkind.”😭
“Wld u let yr own fmly live in thes conds?”... “It’s unethical. It’s irresponsible. It’s unkind.”😭
‘Hell on earth’: immigrants held in new California detention facility beg for help
Six people at a remote Ice facility in California City describe medical problems and filthy conditions, which operator CoreCivic denies
www.theguardian.com
September 27, 2025 at 3:39 PM
Everybody can reply
7 reposts
10 likes
1 saves
"sample machine go brms"
September 19, 2025 at 1:14 PM
Everybody can reply
8 likes
brms. No, I won’t reveal the correct pronunciation.
September 19, 2025 at 9:42 AM
Everybody can reply
19 likes
September 7, 2025 at 2:55 AM
Everybody can reply
4 likes
I'm glad you like it!
My ultimate goal is to be able to get all the way from stan/brms/whatever to any visualization without breaking out of the probabilistic programming abstraction. It's not 100% (probably never will be) but it's getting there
My ultimate goal is to be able to get all the way from stan/brms/whatever to any visualization without breaking out of the probabilistic programming abstraction. It's not 100% (probably never will be) but it's getting there
May 9, 2024 at 7:00 PM
Everybody can reply
1 likes
🌍 Interested in ecological time series? Go beyond ARIMA!
Join us to learn how to wrangle, visualize, and analyze ecological time series using the {mvgam} and {brms} packages to generate accurate forecasts.
👨🍳 @njclark.bsky.social
📅 24-28 March
www.physalia-courses.org/courses-work...
Join us to learn how to wrangle, visualize, and analyze ecological time series using the {mvgam} and {brms} packages to generate accurate forecasts.
👨🍳 @njclark.bsky.social
📅 24-28 March
www.physalia-courses.org/courses-work...
time series analysis and forecasting in R
24-28 March 2025 To foster international participation, this course will be held online
www.physalia-courses.org
March 6, 2025 at 8:09 AM
Everybody can reply
4 reposts
1 quotes
12 likes
Would you like to use Bayesian methods, but you worry about how to set priors & the "subjective" choices involved? In this #brms video tutorial @ChrisMMCox and I show how to rigorously build your Bayesian analysis w a focus on motivated priors....
November 13, 2024 at 10:46 PM
Everybody can reply