I am doing an extreme value theory analysis so I built a model with brms and the gen_extreme_value family. Specifically, I am using a truncated response distribution brm(y | trunc(lb = 0.0) ~ ..., family = gen_extreme_value(), .... The sampling works but then I get the error Error: posterior_epred values on the respone scale not yet implemented for truncated 'gen_extreme_value' models. during postprocessing. Is this something that could be implemented in tidybayes or shall this be fixed in brms directly?
Ah yeah, that's a brms issue unfortunately --- can't be fixed on the tidybayes side, which depends on brms::posterior_epred() having an implementation for the model in question. Sorry I can't be of more help!
I am doing an extreme value theory analysis so I built a model with brms and the
gen_extreme_value
family. Specifically, I am using a truncated response distributionbrm(y | trunc(lb = 0.0) ~ ..., family = gen_extreme_value(), ...
. The sampling works but then I get the errorError: posterior_epred values on the respone scale not yet implemented for truncated 'gen_extreme_value' models.
during postprocessing. Is this something that could be implemented in tidybayes or shall this be fixed in brms directly?