Create a Generic Model Specification — create_model

Canonical constructor for custom analysis/model specifications used by rMVPA's S3 dispatch.

Usage

create_model_spec(
  name,
  dataset,
  design,
  return_predictions = FALSE,
  compute_performance = FALSE,
  tune_reps = FALSE,
  ...
)

name: A non-empty character scalar indicating the model class name.
dataset: An object inheriting from mvpa_dataset.
design: A design object, typically mvpa_design or a specialized *_design class.
return_predictions: Logical scalar indicating whether row-wise predictions should be retained.
compute_performance: Logical scalar indicating whether performance metrics should be computed during ROI iteration.
tune_reps: Tuning replication control used by model implementations.
...: Additional named fields stored in the model specification. If has_test_set is not provided, it is inferred from dataset$test_data and design$y_test.

A list with class c(name, "model_spec", "list").