def test_permutation_importance_correlated_feature_regression_pandas(n_jobs):
pd = pytest.importorskip("pandas")
# Make sure that feature highly correlated to the target have a higher
# importance
rng = np.random.RandomState(42)
n_repeats = 5
dataset = load_iris()
X, y = dataset.data, dataset.target
y_with_little_noise = (y +
rng.normal(scale=0.001, size=y.shape[0])).reshape(
-1, 1)
# Adds feature correlated with y as the last column
X = pd.DataFrame(X, columns=dataset.feature_names)
X['correlated_feature'] = y_with_little_noise
clf = RandomForestClassifier(n_estimators=10, random_state=42)
clf.fit(X, y)
result = permutation_importance(clf,
X,
y,
n_repeats=n_repeats,
random_state=rng,
n_jobs=n_jobs)
assert result.importances.shape == (X.shape[1], n_repeats)
# the correlated feature with y was added as the last column and should
# have the highest importance
assert np.all(result.importances_mean[-1] > result.importances_mean[:-1])
def test_rfe_estimator_tags():
rfe = RFE(SVC(kernel='linear'))
assert rfe._estimator_type == "classifier"
# make sure that cross-validation is stratified
iris = load_iris()
score = cross_val_score(rfe, iris.data, iris.target)
assert score.min() > .7
def test_rfe():
generator = check_random_state(0)
iris = load_iris()
X = np.c_[iris.data, generator.normal(size=(len(iris.data), 6))]
X_sparse = sparse.csr_matrix(X)
y = iris.target
# dense model
clf = SVC(kernel="linear")
rfe = RFE(estimator=clf, n_features_to_select=4, step=0.1)
rfe.fit(X, y)
X_r = rfe.transform(X)
clf.fit(X_r, y)
assert len(rfe.ranking_) == X.shape[1]
# sparse model
clf_sparse = SVC(kernel="linear")
rfe_sparse = RFE(estimator=clf_sparse, n_features_to_select=4, step=0.1)
rfe_sparse.fit(X_sparse, y)
X_r_sparse = rfe_sparse.transform(X_sparse)
assert X_r.shape == iris.data.shape
assert_array_almost_equal(X_r[:10], iris.data[:10])
assert_array_almost_equal(rfe.predict(X), clf.predict(iris.data))
assert rfe.score(X, y) == clf.score(iris.data, iris.target)
assert_array_almost_equal(X_r, X_r_sparse.toarray())
def test_base():
# Check BaseEnsemble methods.
ensemble = BaggingClassifier(base_estimator=Perceptron(random_state=None),
n_estimators=3)
iris = load_iris()
ensemble.fit(iris.data, iris.target)
ensemble.estimators_ = [] # empty the list and create estimators manually
ensemble._make_estimator()
random_state = np.random.RandomState(3)
ensemble._make_estimator(random_state=random_state)
ensemble._make_estimator(random_state=random_state)
ensemble._make_estimator(append=False)
assert 3 == len(ensemble)
assert 3 == len(ensemble.estimators_)
assert isinstance(ensemble[0], Perceptron)
assert ensemble[0].random_state is None
assert isinstance(ensemble[1].random_state, int)
assert isinstance(ensemble[2].random_state, int)
assert ensemble[1].random_state != ensemble[2].random_state
np_int_ensemble = BaggingClassifier(base_estimator=Perceptron(),
n_estimators=np.int32(3))
np_int_ensemble.fit(iris.data, iris.target)
def test_feature_union_weights():
# test feature union with transformer weights
iris = load_iris()
X = iris.data
y = iris.target
pca = PCA(n_components=2, svd_solver='randomized', random_state=0)
select = SelectKBest(k=1)
# test using fit followed by transform
fs = FeatureUnion([("pca", pca), ("select", select)],
transformer_weights={"pca": 10})
fs.fit(X, y)
X_transformed = fs.transform(X)
# test using fit_transform
fs = FeatureUnion([("pca", pca), ("select", select)],
transformer_weights={"pca": 10})
X_fit_transformed = fs.fit_transform(X, y)
# test it works with transformers missing fit_transform
fs = FeatureUnion([("mock", Transf()), ("pca", pca), ("select", select)],
transformer_weights={"mock": 10})
X_fit_transformed_wo_method = fs.fit_transform(X, y)
# check against expected result
# We use a different pca object to control the random_state stream
assert_array_almost_equal(X_transformed[:, :-1], 10 * pca.fit_transform(X))
assert_array_equal(X_transformed[:, -1],
select.fit_transform(X, y).ravel())
assert_array_almost_equal(X_fit_transformed[:, :-1],
10 * pca.fit_transform(X))
assert_array_equal(X_fit_transformed[:, -1],
select.fit_transform(X, y).ravel())
assert X_fit_transformed_wo_method.shape == (X.shape[0], 7)
def test_graphical_lasso_iris_singular():
# Small subset of rows to test the rank-deficient case
# Need to choose samples such that none of the variances are zero
indices = np.arange(10, 13)
# Hard-coded solution from R glasso package for alpha=0.01
cov_R = np.array([
[0.08, 0.056666662595, 0.00229729713223, 0.00153153142149],
[0.056666662595, 0.082222222222, 0.00333333333333, 0.00222222222222],
[0.002297297132, 0.003333333333, 0.00666666666667, 0.00009009009009],
[0.001531531421, 0.002222222222, 0.00009009009009, 0.00222222222222]
])
icov_R = np.array([
[24.42244057, -16.831679593, 0.0, 0.0],
[-16.83168201, 24.351841681, -6.206896552, -12.5],
[0.0, -6.206896171, 153.103448276, 0.0],
[0.0, -12.499999143, 0.0, 462.5]
])
X = datasets.load_iris().data[indices, :]
emp_cov = empirical_covariance(X)
for method in ('cd', 'lars'):
cov, icov = graphical_lasso(emp_cov, alpha=0.01, return_costs=False,
mode=method)
assert_array_almost_equal(cov, cov_R, decimal=5)
assert_array_almost_equal(icov, icov_R, decimal=5)
def test_pipeline_methods_preprocessing_svm():
# Test the various methods of the pipeline (preprocessing + svm).
iris = load_iris()
X = iris.data
y = iris.target
n_samples = X.shape[0]
n_classes = len(np.unique(y))
scaler = StandardScaler()
pca = PCA(n_components=2, svd_solver='randomized', whiten=True)
clf = SVC(probability=True, random_state=0, decision_function_shape='ovr')
for preprocessing in [scaler, pca]:
pipe = Pipeline([('preprocess', preprocessing), ('svc', clf)])
pipe.fit(X, y)
# check shapes of various prediction functions
predict = pipe.predict(X)
assert predict.shape == (n_samples,)
proba = pipe.predict_proba(X)
assert proba.shape == (n_samples, n_classes)
log_proba = pipe.predict_log_proba(X)
assert log_proba.shape == (n_samples, n_classes)
decision_function = pipe.decision_function(X)
assert decision_function.shape == (n_samples, n_classes)
pipe.score(X, y)
def test_check_estimator_clones():
# check that check_estimator doesn't modify the estimator it receives
from mrex.datasets import load_iris
iris = load_iris()
for Estimator in [
GaussianMixture, LinearRegression, RandomForestClassifier, NMF,
SGDClassifier, MiniBatchKMeans
]:
with ignore_warnings(category=(FutureWarning, DeprecationWarning)):
# when 'est = SGDClassifier()'
est = Estimator()
set_checking_parameters(est)
set_random_state(est)
# without fitting
old_hash = joblib.hash(est)
check_estimator(est)
assert old_hash == joblib.hash(est)
with ignore_warnings(category=(FutureWarning, DeprecationWarning)):
# when 'est = SGDClassifier()'
est = Estimator()
set_checking_parameters(est)
set_random_state(est)
# with fitting
est.fit(iris.data + 10, iris.target)
old_hash = joblib.hash(est)
check_estimator(est)
assert old_hash == joblib.hash(est)
def test_pickle_version_warning_is_issued_upon_different_version():
iris = datasets.load_iris()
tree = TreeBadVersion().fit(iris.data, iris.target)
tree_pickle_other = pickle.dumps(tree)
message = pickle_error_message.format(estimator="TreeBadVersion",
old_version="something",
current_version=mrex.__version__)
assert_warns_message(UserWarning, message, pickle.loads, tree_pickle_other)
def test_base_zero_n_estimators():
# Check that instantiating a BaseEnsemble with n_estimators<=0 raises
# a ValueError.
ensemble = BaggingClassifier(base_estimator=Perceptron(), n_estimators=0)
iris = load_iris()
assert_raise_message(ValueError,
"n_estimators must be greater than zero, got 0.",
ensemble.fit, iris.data, iris.target)
def test_non_encoded_labels():
dataset = datasets.load_iris()
X = dataset.data
labels = dataset.target
assert (silhouette_score(X,
labels * 2 + 10) == silhouette_score(X, labels))
assert_array_equal(silhouette_samples(X, labels * 2 + 10),
silhouette_samples(X, labels))
def test_gnb_naive_bayes_scale_invariance():
# Scaling the data should not change the prediction results
iris = load_iris()
X, y = iris.data, iris.target
labels = [
GaussianNB().fit(f * X, y).predict(f * X) for f in [1E-10, 1, 1E10]
]
assert_array_equal(labels[0], labels[1])
assert_array_equal(labels[1], labels[2])
def test_load_iris():
res = load_iris()
assert res.data.shape == (150, 4)
assert res.target.size == 150
assert res.target_names.size == 3
assert res.DESCR
assert os.path.exists(res.filename)
# test return_X_y option
check_return_X_y(res, partial(load_iris))
def test_pickle_version_no_warning_is_issued_with_non_mrex_estimator():
iris = datasets.load_iris()
tree = TreeNoVersion().fit(iris.data, iris.target)
tree_pickle_noversion = pickle.dumps(tree)
try:
module_backup = TreeNoVersion.__module__
TreeNoVersion.__module__ = "notmrex"
assert_no_warnings(pickle.loads, tree_pickle_noversion)
finally:
TreeNoVersion.__module__ = module_backup
def test_pickle_version_warning_is_not_raised_with_matching_version():
iris = datasets.load_iris()
tree = DecisionTreeClassifier().fit(iris.data, iris.target)
tree_pickle = pickle.dumps(tree)
assert b"version" in tree_pickle
tree_restored = assert_no_warnings(pickle.loads, tree_pickle)
# test that we can predict with the restored decision tree classifier
score_of_original = tree.score(iris.data, iris.target)
score_of_restored = tree_restored.score(iris.data, iris.target)
assert score_of_original == score_of_restored
def test_rfecv_mockclassifier():
generator = check_random_state(0)
iris = load_iris()
X = np.c_[iris.data, generator.normal(size=(len(iris.data), 6))]
y = list(iris.target) # regression test: list should be supported
# Test using the score function
rfecv = RFECV(estimator=MockClassifier(), step=1)
rfecv.fit(X, y)
# non-regression test for missing worst feature:
assert len(rfecv.grid_scores_) == X.shape[1]
assert len(rfecv.ranking_) == X.shape[1]
def test_base_not_int_n_estimators():
# Check that instantiating a BaseEnsemble with a string as n_estimators
# raises a ValueError demanding n_estimators to be supplied as an integer.
string_ensemble = BaggingClassifier(base_estimator=Perceptron(),
n_estimators='3')
iris = load_iris()
assert_raise_message(ValueError, "n_estimators must be an integer",
string_ensemble.fit, iris.data, iris.target)
float_ensemble = BaggingClassifier(base_estimator=Perceptron(),
n_estimators=3.0)
assert_raise_message(ValueError, "n_estimators must be an integer",
float_ensemble.fit, iris.data, iris.target)
def test_pipeline_fit_transform():
# Test whether pipeline works with a transformer missing fit_transform
iris = load_iris()
X = iris.data
y = iris.target
transf = Transf()
pipeline = Pipeline([('mock', transf)])
# test fit_transform:
X_trans = pipeline.fit_transform(X, y)
X_trans2 = transf.fit(X, y).transform(X)
assert_array_almost_equal(X_trans, X_trans2)
def test_pickle_version_warning_is_issued_when_no_version_info_in_pickle():
iris = datasets.load_iris()
# TreeNoVersion has no getstate, like pre-0.18
tree = TreeNoVersion().fit(iris.data, iris.target)
tree_pickle_noversion = pickle.dumps(tree)
assert b"version" not in tree_pickle_noversion
message = pickle_error_message.format(estimator="TreeNoVersion",
old_version="pre-0.18",
current_version=mrex.__version__)
# check we got the warning about using pre-0.18 pickle
assert_warns_message(UserWarning, message, pickle.loads,
tree_pickle_noversion)
def test_pipeline_wrong_memory():
# Test that an error is raised when memory is not a string or a Memory
# instance
iris = load_iris()
X = iris.data
y = iris.target
# Define memory as an integer
memory = 1
cached_pipe = Pipeline([('transf', DummyTransf()),
('svc', SVC())], memory=memory)
assert_raises_regex(ValueError, "'memory' should be None, a string or"
" have the same interface as joblib.Memory."
" Got memory='1' instead.", cached_pipe.fit, X, y)
def test_classes_property():
iris = load_iris()
X = iris.data
y = iris.target
reg = make_pipeline(SelectKBest(k=1), LinearRegression())
reg.fit(X, y)
assert_raises(AttributeError, getattr, reg, "classes_")
clf = make_pipeline(SelectKBest(k=1), LogisticRegression(random_state=0))
assert_raises(AttributeError, getattr, clf, "classes_")
clf.fit(X, y)
assert_array_equal(clf.classes_, np.unique(y))
def test_rfe_mockclassifier():
generator = check_random_state(0)
iris = load_iris()
X = np.c_[iris.data, generator.normal(size=(len(iris.data), 6))]
y = iris.target
# dense model
clf = MockClassifier()
rfe = RFE(estimator=clf, n_features_to_select=4, step=0.1)
rfe.fit(X, y)
X_r = rfe.transform(X)
clf.fit(X_r, y)
assert len(rfe.ranking_) == X.shape[1]
assert X_r.shape == iris.data.shape
def test_pipeline_methods_pca_svm():
# Test the various methods of the pipeline (pca + svm).
iris = load_iris()
X = iris.data
y = iris.target
# Test with PCA + SVC
clf = SVC(probability=True, random_state=0)
pca = PCA(svd_solver='full', n_components='mle', whiten=True)
pipe = Pipeline([('pca', pca), ('svc', clf)])
pipe.fit(X, y)
pipe.predict(X)
pipe.predict_proba(X)
pipe.predict_log_proba(X)
pipe.score(X, y)
def test_pipeline_methods_anova():
# Test the various methods of the pipeline (anova).
iris = load_iris()
X = iris.data
y = iris.target
# Test with Anova + LogisticRegression
clf = LogisticRegression()
filter1 = SelectKBest(f_classif, k=2)
pipe = Pipeline([('anova', filter1), ('logistic', clf)])
pipe.fit(X, y)
pipe.predict(X)
pipe.predict_proba(X)
pipe.predict_log_proba(X)
pipe.score(X, y)
def test_ovo_partial_fit_predict():
temp = datasets.load_iris()
X, y = temp.data, temp.target
ovo1 = OneVsOneClassifier(MultinomialNB())
ovo1.partial_fit(X[:100], y[:100], np.unique(y))
ovo1.partial_fit(X[100:], y[100:])
pred1 = ovo1.predict(X)
ovo2 = OneVsOneClassifier(MultinomialNB())
ovo2.fit(X, y)
pred2 = ovo2.predict(X)
assert len(ovo1.estimators_) == n_classes * (n_classes - 1) / 2
assert np.mean(y == pred1) > 0.65
assert_almost_equal(pred1, pred2)
# Test when mini-batches have binary target classes
ovo1 = OneVsOneClassifier(MultinomialNB())
ovo1.partial_fit(X[:60], y[:60], np.unique(y))
ovo1.partial_fit(X[60:], y[60:])
pred1 = ovo1.predict(X)
ovo2 = OneVsOneClassifier(MultinomialNB())
pred2 = ovo2.fit(X, y).predict(X)
assert_almost_equal(pred1, pred2)
assert len(ovo1.estimators_) == len(np.unique(y))
assert np.mean(y == pred1) > 0.65
ovo = OneVsOneClassifier(MultinomialNB())
X = np.random.rand(14, 2)
y = [1, 1, 2, 3, 3, 0, 0, 4, 4, 4, 4, 4, 2, 2]
ovo.partial_fit(X[:7], y[:7], [0, 1, 2, 3, 4])
ovo.partial_fit(X[7:], y[7:])
pred = ovo.predict(X)
ovo2 = OneVsOneClassifier(MultinomialNB())
pred2 = ovo2.fit(X, y).predict(X)
assert_almost_equal(pred, pred2)
# raises error when mini-batch does not have classes from all_classes
ovo = OneVsOneClassifier(MultinomialNB())
error_y = [0, 1, 2, 3, 4, 5, 2]
message_re = escape("Mini-batch contains {0} while "
"it must be subset of {1}".format(
np.unique(error_y), np.unique(y)))
assert_raises_regexp(ValueError, message_re, ovo.partial_fit, X[:7],
error_y, np.unique(y))
# test partial_fit only exists if estimator has it:
ovr = OneVsOneClassifier(SVC())
assert not hasattr(ovr, "partial_fit")
def test_rfe_cv_groups():
generator = check_random_state(0)
iris = load_iris()
number_groups = 4
groups = np.floor(np.linspace(0, number_groups, len(iris.target)))
X = iris.data
y = (iris.target > 0).astype(int)
est_groups = RFECV(
estimator=RandomForestClassifier(random_state=generator),
step=1,
scoring='accuracy',
cv=GroupKFold(n_splits=2))
est_groups.fit(X, y, groups=groups)
assert est_groups.n_features_ > 0
def test_graph_lasso_2D():
# Hard-coded solution from Python skggm package
# obtained by calling `quic(emp_cov, lam=.1, tol=1e-8)`
cov_skggm = np.array([[3.09550269, 1.186972],
[1.186972, 0.57713289]])
icov_skggm = np.array([[1.52836773, -3.14334831],
[-3.14334831, 8.19753385]])
X = datasets.load_iris().data[:, 2:]
emp_cov = empirical_covariance(X)
for method in ('cd', 'lars'):
cov, icov = graphical_lasso(emp_cov, alpha=.1, return_costs=False,
mode=method)
assert_array_almost_equal(cov, cov_skggm)
assert_array_almost_equal(icov, icov_skggm)
def test_pipeline_score_samples_pca_lof():
iris = load_iris()
X = iris.data
# Test that the score_samples method is implemented on a pipeline.
# Test that the score_samples method on pipeline yields same results as
# applying transform and score_samples steps separately.
pca = PCA(svd_solver='full', n_components='mle', whiten=True)
lof = LocalOutlierFactor(novelty=True)
pipe = Pipeline([('pca', pca), ('lof', lof)])
pipe.fit(X)
# Check the shapes
assert pipe.score_samples(X).shape == (X.shape[0],)
# Check the values
lof.fit(pca.fit_transform(X))
assert_allclose(pipe.score_samples(X), lof.score_samples(pca.transform(X)))
def test_rfe_cv_n_jobs():
generator = check_random_state(0)
iris = load_iris()
X = np.c_[iris.data, generator.normal(size=(len(iris.data), 6))]
y = iris.target
rfecv = RFECV(estimator=SVC(kernel='linear'))
rfecv.fit(X, y)
rfecv_ranking = rfecv.ranking_
rfecv_grid_scores = rfecv.grid_scores_
rfecv.set_params(n_jobs=2)
rfecv.fit(X, y)
assert_array_almost_equal(rfecv.ranking_, rfecv_ranking)
assert_array_almost_equal(rfecv.grid_scores_, rfecv_grid_scores)
def test_feature_union():
# basic sanity check for feature union
iris = load_iris()
X = iris.data
X -= X.mean(axis=0)
y = iris.target
svd = TruncatedSVD(n_components=2, random_state=0)
select = SelectKBest(k=1)
fs = FeatureUnion([("svd", svd), ("select", select)])
fs.fit(X, y)
X_transformed = fs.transform(X)
assert X_transformed.shape == (X.shape[0], 3)
# check if it does the expected thing
assert_array_almost_equal(X_transformed[:, :-1], svd.fit_transform(X))
assert_array_equal(X_transformed[:, -1],
select.fit_transform(X, y).ravel())
# test if it also works for sparse input
# We use a different svd object to control the random_state stream
fs = FeatureUnion([("svd", svd), ("select", select)])
X_sp = sparse.csr_matrix(X)
X_sp_transformed = fs.fit_transform(X_sp, y)
assert_array_almost_equal(X_transformed, X_sp_transformed.toarray())
# Test clone
fs2 = assert_no_warnings(clone, fs)
assert fs.transformer_list[0][1] is not fs2.transformer_list[0][1]
# test setting parameters
fs.set_params(select__k=2)
assert fs.fit_transform(X, y).shape == (X.shape[0], 4)
# test it works with transformers missing fit_transform
fs = FeatureUnion([("mock", Transf()), ("svd", svd), ("select", select)])
X_transformed = fs.fit_transform(X, y)
assert X_transformed.shape == (X.shape[0], 8)
# test error if some elements do not support transform
assert_raises_regex(TypeError,
'All estimators should implement fit and '
'transform.*\\bNoTrans\\b',
FeatureUnion,
[("transform", Transf()), ("no_transform", NoTrans())])
# test that init accepts tuples
fs = FeatureUnion((("svd", svd), ("select", select)))
fs.fit(X, y)
