Figure 4

Boxplots of scaled permutation-based VIMs using random forest. First column: scaled VIMs of variables under H_A using: 1^st row: mtry = 1, 2^nd row: mtry = 3, 3^rd row: mtry = 8. Second column VIMs of variables under H₀. V = number of predictor (x_i) in the model.