Calulate actual results for every pair of models

Calulate actual results for every pair of models

calculate_actual_wins(
  results,
  decreasing_metric = TRUE,
  compare_in_split,
  aggregate = TRUE
)

Arguments

results	raw results
decreasing_metric	if used metric is decreasing
compare_in_split	if compare models and parameters only in fhe same fold
aggregate	if results should be aggregated for every pair of models and hyperparameters. Otherwise output will have many rows with binary response (acccording to amount of splitsin crossvalidation) for every pair of models