Fig 15. Critical difference diagram for ensemble models by comparing the four classifiers on different combinations of features; (a) Balanced Accuracy scores with α = 0.05, the critical difference is 1.77; (b) F1 with α = 0.1, the critical difference is 1.58; (c) AUC with α = 0.05, the critical difference is 1.77.

Groups of classifiers not significantly different are connected.