Learn R Programming

brms (version 2.16.1)

get_dpar: Draws of a Distributional Parameter

Description

Get draws of a distributional parameter from a brmsprep or mvbrmsprep object. This function is primarily useful when developing custom families or packages depending on brms. This function lets callers easily handle both the case when the distributional parameter is predicted directly, via a (non-)linear predictor or fixed to a constant. See the vignette vignette("brms_customfamilies") for an example use case.

Usage

get_dpar(prep, dpar, i = NULL, ilink = NULL)

Arguments

prep

A 'brmsprep' or 'mvbrmsprep' object created by prepare_predictions.

dpar

Name of the distributional parameter.

i

The observation numbers for which predictions shall be extracted. If NULL (the default), all observation will be extracted. Ignored if dpar is not predicted.

ilink

Should the inverse link function be applied? If NULL (the default), the value is chosen internally. In particular, ilink is TRUE by default for custom families.

Value

If the parameter is predicted and i is NULL or length(i) > 1, an S x N matrix. If the parameter it not predicted or length(i) == 1, a vector of length S. Here S is the number of draws and N is the number of observations or length of i if specified.

Examples

Run this code
# NOT RUN {
posterior_predict_my_dist <- function(i, prep, ...) {
  mu <- brms::get_dpar(prep, "mu", i = i)
  mypar <- brms::get_dpar(prep, "mypar", i = i)
  my_rng(mu, mypar)
}
# }
# NOT RUN {
# }

Run the code above in your browser using DataLab