Skip to contents

Diagnose GAM and BAM models

Source: R/diagnose_gazepoint_gamm.R
diagnose_gazepoint_gamm.Rd

Run a compact diagnostics bundle for mgcv GAM/BAM models used in gp3tools workflows. The function combines convergence, basis-dimension checks, overdispersion checks, and optional DHARMa simulation-based residual diagnostics.

Usage

diagnose_gazepoint_gamm(
  model,
  model_name = NULL,
  check_convergence = TRUE,
  check_basis = TRUE,
  check_overdispersion = TRUE,
  use_dharma = FALSE,
  dharma_simulations = 250,
  seed = 123
)

Arguments

model

A fitted GAM/BAM object, a gp3tools fit object containing $model, or a named list of fitted model objects.

model_name

Optional model label used in returned tables.

check_convergence

Logical. If TRUE, run convergence diagnostics.

check_basis

Logical. If TRUE, run mgcv::k.check() basis-dimension diagnostics when available.

check_overdispersion

Logical. If TRUE, run overdispersion diagnostics when meaningful for the model family.

use_dharma

Logical. If TRUE, try to run optional DHARMa diagnostics.

dharma_simulations

Number of DHARMa simulations.

seed

Random seed used before DHARMa simulation.

Value

A list with overview, convergence, basis, overdispersion, DHARMa diagnostics, and settings.

Details

The function accepts raw mgcv::gam() / mgcv::bam() model objects, gp3tools fit objects containing a $model element, or a named list of fitted model objects.