R/JAGS-formula-random.R
random_group_component_names.Rdrandom_group_component_names() returns the names of the
grouping variables of a compiled random-effect term.
random_group_prediction_keys() assigns each row of a data frame the key
of its group, so that the rows of new data can be matched to the fitted
groups of the term (or to groups that were not fitted) when random effects
are predicted or marginalized.
random_group_component_names(term)
random_group_prediction_keys(term, data)random_group_component_names() returns a character vector of
the names of the grouping variables of term.
random_group_prediction_keys() returns a character vector
with one key per row of data: the key of the fitted group for rows in
a fitted group of term, and a distinct key for rows in a new group.
A compiled random-effect term is an element of the
random_effects of a formula design: the design returned by
JAGS_formula() (JAGS_formula(...)$formula_design$random_effects[[i]])
or the design stored with a fit or with draws
(JAGS_formula_design(fit, parameter)$random_effects[[i]]). Pass the
element unchanged; its fields are not part of the public interface.
An interaction (g1:g2) or nested (g1/g2) grouping expression
has one grouping variable per component, returned in the order of the
fitted groups.
The keys of random_group_prediction_keys() are character strings
whose format is not part of the interface. Rows with the same grouping values
share a key; a row whose grouping values identify a fitted group of the term
has the key of that group, and a row with values that are not a fitted group
has a key of its own that differs from the keys of all fitted groups.
A grouping value is matched to the fitted levels by its exact value; a value
that matches no fitted level but equals the label (or the displayed value) of
exactly one fitted level takes that level, and a label shared by several
fitted levels stops with an error. A term without its fitted grouping
ownership (a design stored by an earlier version) stops with an error of
class BayesTools_refit_required. A grouping variable that is missing
in data or contains missing values stops with an error.