default_prior() documentation

[ASKurz]

In the documentation for default_prior() we read:

default_prior is a generic function that can be used to get default priors for Bayesian models.

This seems a little misleading. In the example below where a model uses non-default priors, default_prior() returns the same output as get_prior(). As the documentation currently reads, I would have expected default_prior() to return what the defaults would have been had I not used the prior argument within brm(), and I would have found it very helpful if it had.

fit1 <- brm(
  count ~ zAge + zBase * Trt + (1|patient) + (1|obs),
  data = epilepsy, 
  family = poisson(),
  prior = c(
    prior(normal(0, 1), class = Intercept),
    prior(normal(0, 1), class = b),
    prior(exponential(1), class = sd)),
  cores = 4, iter = 500)

default_prior(fit1)

          prior     class       coef   group resp dpar nlpar lb ub       source
   normal(0, 1)         b                                                  user
   normal(0, 1)         b       Trt1                               (vectorized)
   normal(0, 1)         b       zAge                               (vectorized)
   normal(0, 1)         b      zBase                               (vectorized)
   normal(0, 1)         b zBase:Trt1                               (vectorized)
   normal(0, 1) Intercept                                                  user
 exponential(1)        sd                                     0            user
 exponential(1)        sd                obs                  0    (vectorized)
 exponential(1)        sd  Intercept     obs                  0    (vectorized)
 exponential(1)        sd            patient                  0    (vectorized)
 exponential(1)        sd  Intercept patient                  0    (vectorized)

The weaker part of my request is more clarity in the documentation. But the stronger part of my request is it would be super cool if default_prior() returned what the default prior settings would have been for a model fit with non-default priors.

[paul-buerkner]

I like the idea. Will see what to best make out of it. :-)