objectif

Mesurer l’importance en permutant une feature.

code minimal

from sklearn.inspection import permutation_importance
from sklearn.ensemble import RandomForestClassifier
from sklearn.datasets import load_iris

X, y = load_iris(return_X_y=True)
clf = RandomForestClassifier(random_state=0).fit(X, y)
r = permutation_importance(clf, X, y, n_repeats=5, random_state=0)
print((r.importances_mean >= 0.0).all())

utilisation

import numpy as np
order = np.argsort(r.importances_mean)[::-1]
print(order[:2].tolist())

variante(s) utile(s)

from sklearn.ensemble import RandomForestRegressor
print(hasattr(RandomForestRegressor(), "fit"))

notes

• Calculer sur un set de validation pour éviter le biais.