I ran hyperband once with holdout as resampling, and once with 5-fold cross-validation. For the cross-validation, the wrong individuals "survive", which is exemplarily visualized for a single bracket:
Reason for this is this line of code, which should be instance$bmr$aggregate.
I ran hyperband once with holdout as resampling, and once with 5-fold cross-validation. For the cross-validation, the wrong individuals "survive", which is exemplarily visualized for a single bracket:
Reason for this is this line of code, which should be
instance$bmr$aggregate.