Greetings.
I would like to ask you for help in compiling a custom model. I am interested in a beta regression model for the case of ordinal scale (https://cran.rstudio.com/web/packages/DBR/vignettes/DBR.pdf). As I understood, it is an alternative to ordbetareg(), beta_binomial(), cumulative(link = ‘probit’) models. However, I lack knowledge how to translate it from formula language to syntax.
HI, @comisara and welcome to the Stan forums.
You could definitely implement this model in Stan from the paper. I’m not sure if brms is expressive enough.
If you want to go off the (0, 1) scale, you can use an ordinal logit. That should be supported in brms and it’s easier to extend to a hierarchical model on the log odds scale.
1 Like