Compare Because Models — because

A unified function to either (1) compare previously fitted models, or (2) run multiple model specifications in parallel and then compare them.

Usage

because_compare(
  ...,
  model_specs = NULL,
  data = NULL,
  tree = NULL,
  n.cores = 1,
  cl = NULL,
  sort = TRUE
)

Arguments

...: For comparing fitted models: individual fitted model objects of class "because". For running models: additional arguments passed to because (e.g., n.iter).
model_specs: A named list of model specifications to run (Mode 2). Each element should be a list containing arguments for because. Alternatively, this argument can accept the first fitted model object (Mode 1).
data: The dataset (required for Mode 2). Alternatively, the second fitted model object (Mode 1).
tree: The phylogenetic tree (optional for Mode 2). Alternatively, the third fitted model object (Mode 1).
n.cores: Number of cores for parallel execution (Mode 2). Default is 1.
cl: Optional cluster object (Mode 2).
sort: Logical. If TRUE (default), sort comparison table by WAIC.

Value

If comparing fitted models: A class "because_comparison" object (data frame) with WAIC rankings.

If running models: A list containing:

results: List of fitted model objects.
comparison: The comparison data frame.

Details

Mode 1: Compare Fitted Models Call because_compare(fit1, fit2) or because_compare(models = list(fit1, fit2)). Extracts WAIC (with SE) from each model and ranks them.

Mode 2: Run and Compare Call because_compare(model_specs = list(m1=..., m2=...), data=data, tree=tree). This runs the models in parallel and returns the comparison.

Examples

if (FALSE) { # \dontrun{
  # Mode 1: Compare existing fits
  because_compare(fit1, fit2)

  # Mode 2: Run and compare
  specs <- list(m1 = list(equations = list(Y ~ X)), m2 = list(equations = list(Y ~ X + Z)))
  res <- because_compare(specs, data = df, tree = tr, n.cores = 2)
  print(res$comparison)
} # }