Influence diagnostics for brms models — influ.brmsfit • influ2

Population-level posterior coefficients are projected directly into the compact focus-by-term contrasts. The method therefore never constructs an observation-by-draw-by-term array. Joint posterior dependence is preserved while calculating the requested diagnostics. Lognormal models currently require constant `sigma`; varying scale needs a joint location-and-scale calculation before arithmetic-mean ratios can be reported. The `mu` link must be identity because brms parameterises the log-location, not the arithmetic mean. Formula offsets and response `rate()` additions have the same restrictions as the GLM adapter.

Usage

# S3 method for class 'brmsfit'
influ(
  model,
  focus,
  data = NULL,
  weights = NULL,
  reference_data = NULL,
  reference_weights = NULL,
  uncertainty = "auto",
  retain = "summary",
  probs = c(0.025, 0.975),
  ndraws = NULL,
  seed = NULL,
  draws_path = NULL,
  keep_model = FALSE,
  ...
)

Arguments

model: A fitted object from [brms::brm()].
focus: Name of the focus variable, usually year.
data: Optional model data. The model frame is used by default.
weights: Optional numeric observation weights or the name of a weight column in `data`.
reference_data: Optional prediction grid defining the common standardisation distribution. By default the observed data are used.
reference_weights: Optional numeric weights, or a column name, for `reference_data`.
uncertainty: Either `"auto"`, `"posterior"`, or `"none"`. The `"none"` mode reduces the posterior to coefficient means before influence calculations.
retain: One of `"summary"`, `"derived_draws"`, or `"disk"`.
probs: Lower and upper interval probabilities.
ndraws: Optional maximum number of posterior draws. `NULL` uses every available draw.
seed: Optional simulation seed.
draws_path: Output path used when `retain = "disk"`.
keep_model: Retain the fitted model inside the diagnostic object.
...: Reserved for future backend options.

Value

An [influ_diag] object.