Extract subject-level posterior draws as an array — extract_subject

The per-draw subject values of every parameter with a group-level effect: the population intercept plus the subject's deviation, summed within each draw, on the link scale. This is what the draws and point estimators of extract_correlations() read.

Usage

extract_subject_draws(fit, group = NULL, ...)

# Default S3 method
extract_subject_draws(fit, group = NULL, ...)

# S3 method for class 'brmsfit'
extract_subject_draws(fit, group = NULL, ...)

Arguments

fit: A brmsfit, and so also a bmmfit. Other classes can provide a method that returns an array of the same shape.
group: The grouping factor to read. NULL uses the fit's only grouping factor and errors if there is more than one.
...: Not used.

Value

A numeric array with the dimensions iteration, chain, id and term, named in its dimnames. term is the parameter name, or <parameter>_<coefficient> (kappa_task1) for a parameter with several group-level coefficients and no intercept, as in extract_estimates(). Parameters the model fixed to a constant (zero posterior variance for every subject) are dropped. The grouping factor is stored in the attribute group.

Examples

if (FALSE) { # \dontrun{
draws <- extract_subject_draws(fit)
dim(draws)
dimnames(draws)$term
} # }