def test_classifier(iris_train):
X, y, feature_names, target_names = iris_train
clf = LogisticRegression().fit(X, y)
assert is_classifier(clf)
perm = PermutationImportance(clf, random_state=42).fit(X, y)
assert is_classifier(perm)
assert (perm.classes_ == [0, 1, 2]).all()
assert np.allclose(clf.predict(X), perm.predict(X))
assert np.allclose(clf.predict_proba(X), perm.predict_proba(X))
assert np.allclose(clf.predict_log_proba(X), perm.predict_log_proba(X))
assert np.allclose(clf.decision_function(X), perm.decision_function(X))
top=None, # show permutation importances for all features
feature_names=feature_names
)
from sklearn.metrics import mean_squared_error, r2_score
# Coefficient of determination r2 for the training set
pipeline_score = permuter.score(X_train_transformed,y_train)
print("Coefficient of determination r2 for the training set.: ", pipeline_score)
# Coefficient of determination r2 for the validation set
pipeline_score = permuter.score(X_val_transformed,y_val)
print("Coefficient of determination r2 for the validation set.: ", pipeline_score)
# The mean squared error
y_pred = permuter.predict(X_val_transformed)
print("Mean squared error: %.2f"% mean_squared_error(y_val, y_pred))
# Thus, Density remains important according to feature permutation than according to feature importance in the Random Fo
# Use importances for feature selection
print('Shape before removing features:', X_train.shape)
# Remove features of 0 importance
zero_importance = 0.0
mask = permuter.feature_importances_ > zero_importance
features1 = X_train.columns[mask]
X_train = X_train[features1]
print('Shape after removing features:', X_train.shape)
# Random forest classifier with 22 features
X_val = X_val[features1]
