Brms generator uses a brms model with sample_prior = "only"
to generate
new datasets.
SBC_generator_brms(
formula,
data,
...,
generate_lp = TRUE,
generate_lp_chunksize = 5000/nrow(data),
out_stan_file = NULL
)
arguments passed to brms::brm
whether to compute the overall log-likelihood of the model as an additional variable. This can be somewhat computationally expensive, but improves sensitivity of the SBC process.
Will determine whether
the chunk size used with future.apply::future_mapply()
. Set quite high
by default as the parallelism only benefits for large individual datasets/number of
simulations.
A filename for the generated Stan code. Useful for debugging and for avoiding unnecessary recompilations.