More vignettes - Githubissues

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

1.29k stars 187 forks source link

More vignettes #155

Closed paul-buerkner closed 10 months ago

paul-buerkner commented 7 years ago

With the implementation of several new features in the upcoming brms 1.3.0 release, some more vignettes are required to explain certain aspects of the package in detail.

[x] Multilevel models (actually write a full paper)
[ ] Prior distributions
[ ] Noise-in-variables models (within the missing values vignette?)
[x] Response time models
[ ] Mixture models
[x] Gaussian processes
[ ] Survival models
[x] Bayes factors
[x] Multivariate models
[x] Missing value imputation

t-barrett commented 7 years ago

It would be great to have a vignette for response time models! Thank you for all of your progress.

ianhussey commented 7 years ago

Seconded. Mixed models for RT data are increasingly commonplace (Balota & Yap, 2011; Lo & Andrews, 2015), but I'm still encountering many researcher degrees of freedom in constructing them in brms. Mixed within-between designs (e.g,. block*condition) cover a large range of paradigms and user cases, such as stroop, IAT, flanker, task switching, etc.

paul-buerkner commented 7 years ago

I agree that a vignette discussing RT models is needed. I wonder though, how such a vignette would lead to a reduction in researchers degrees of freedom? After all, I think it should be about reasonable options and serve as a starting point, which may even increase the researchers degrees of freedom.

paul-buerkner commented 7 years ago

Henrik Singmann has written a blog post about the Wiener Diffusion model (http://singmann.org/wiener-model-analysis-with-brms-part-i/), which might be helpful to many using brms for the analysis of response times.

paoloeusebi commented 5 years ago

It would be great to have something related to multilevel mixture models in the "Advanced Bayesian Multilevel Modeling with the R Package brms" vignette. I don't know if it's possible to estimate such models with brms (see Vermunt 2003. "Multilevel Latent Class Models". Sociological Methodology, 33(1), 213–239.)

paul-buerkner commented 5 years ago

Not in that vignette as it corresponds to an article published in the R journal but of course having a vignette about mixture models would be desireble. If these exact models are possible in brms is another story I have to dig into once I have the time.

paul-buerkner commented 10 months ago

brms is already quite large in term of MB pre-compliation which makes CRAN a bit unhappy. More vignettes will just make it even larger. Given the amount of blog posts and tutorials about brms on the internet, I think it is okay to not add more vignettes to brms itself for the time being. Closing this issue now.