R-User but Stan-Curious

General
,
1

I’ve been heavily using the R ‘brms’ package for a while now and love it. I’ve been a power user of the multivariate distributional models with splines (among other things). My question is if there is value in my transitioning my work over to the proper STAN language? Will I have enhanced capabilities? Will things run faster? Are there other benefits?

Trying to determine if ‘juice is worth the squeeze’. I’d appreciate input!

2

If you find yourself wanting to express a model structure that isn’t practical or intuitive in brms, or if you want your code to more closely reflect the way models are described with math, Stan will give you flexibility and transparency.

3

you can always start with brms and then extend or check for optimizations by using stancode to get the stan that brms is generating for you.