Skip to contents

Utility functions for bru observation model GCPO

Source: R/gcpo.R
bru_obs_gcpo.Rd

Extract and combine INLA::control.gcpo options from bru_obs and bru_obs_list objects, adjusting observation indices for multi-likelihood models where the response vector is the concatenation of all likelihoods.

Usage

bru_obs_control_gcpo(x, ...)

# S3 method for class 'bru_obs'
bru_obs_control_gcpo(
  x,
  index_offset,
  index_length = bru_response_size(x),
  force_weights,
  ...
)

# S3 method for class 'bru_obs_list'
bru_obs_control_gcpo(x, control.gcpo = NULL, ...)

Arguments

x

A bru_obs or bru_obs_list object.

...

Further arguments passed to submethods.

index_offset

integer; offset to add to indices in control.gcpo, equal to the total number of response rows from all preceding likelihoods.

index_length

integer; number of response rows for this likelihood. Defaults to bru_response_size(x).

force_weights

logical; if TRUE, ensure that control.gcpo$weights is populated. Required when any likelihood in a bru_obs_list has non-NULL control.gcpo$weights.

control.gcpo

list of INLA::control.gcpo default options to merge with per-likelihood settings.

Value

  • bru_obs_control_gcpo() returns a list with INLA::control.gcpo options, with predictor/response variable indices unified for multi-observation models. If a bru_obs model has a control.gcpo = NULL argument, an empty list is returned for that model.

See also

bru_obs(), bru_block_gcpo()