model_selection function

Compare fitted models

This function returns a table with several criteria for model comparison.

model_selection(
  ...,
  criteria = c("npar", "LL", "AIC", "BIC"),
  add_form = FALSE
)

Arguments

• ...: One or more objects of class RprobitB_fit.

• criteria: A vector of one or more of the following characters:

  • "npar" for the number of model parameters (see npar),
  • "LL" for the log-likelihood value (see logLik),
  • "AIC" for the AIC value (see AIC),
  • "BIC" for the BIC value (see BIC),
  • "WAIC" for the WAIC value (also shows its standard error sd(WAIC) and the number pWAIC of effective model parameters, see WAIC),
  • "MMLL" for the marginal model log-likelihood,
  • "BF" for the Bayes factor,
  • "pred_acc" for the prediction accuracy (see pred_acc).

• add_form: Set to TRUE to add the model formulas.

Returns

A data frame, criteria in columns, models in rows.

Details

See the vignette on model selection for more details.