Feature Importance in Ensembles

Mean Decrease in Accuracy (MDA)

Permutation importance technique that evaluates the impact of a variable by measuring the performance drop of the model when the values of this variable are randomly permuted on the test set.

← Back