Diagnose a fitted distributional SEM — check

Reports, per node, the family, modelled components, convergence, whether a fixed-effect covariance is available (needed for Wald intervals, d-separation refits, and effect uncertainty), and whether a realized-value sampler exists (needed for distribution-mediated effects). Also lists exogenous variables and warns about anything that will silently degrade a downstream computation.

Usage

check_sem(object, ...)

# S3 method for class 'drm_sem'
check_sem(object, ...)

Arguments

object: A drm_sem object.
...: Unused.

Value

A data frame of per-node diagnostics (class drm_diagnostics).

Examples

if (FALSE) { # \dontrun{
sem <- drm_sem(
  size = drm_node(drmTMB::bf(size ~ temp + habitat, sigma ~ temp),
                  family = stats::gaussian()),
  abundance = drm_node(drmTMB::bf(abundance ~ size + temp, zi ~ habitat),
                       family = drmTMB::nbinom2()),
  data = dat)
check_sem(sem)
} # }