def test_knop():
pool_classifiers, X_dsel, y_dsel, X_test, y_test = setup_classifiers()
kne = KNOP(pool_classifiers, DFP=True)
kne.fit(X_dsel, y_dsel)
assert np.isclose(kne.score(X_test, y_test), 0.9030303030303031)
def test_knop(knne, expected):
pool_classifiers, X_dsel, y_dsel, X_test, y_test = setup_classifiers()
kne = KNOP(pool_classifiers, DFP=True, knne=knne)
kne.fit(X_dsel, y_dsel)
assert np.isclose(kne.score(X_test, y_test), expected)
def test_weights_zero():
knop_test = KNOP(create_pool_classifiers())
knop_test.fit(X_dsel_ex1, y_dsel_ex1)
competences = np.zeros((1, 3))
result = knop_test.select(competences)
assert np.all(result)
def test_select():
knop_test = KNOP(create_pool_classifiers())
competences = np.ones(3)
competences[0] = 0
expected = np.atleast_2d([False, True, True])
selected = knop_test.select(competences)
assert np.array_equal(expected, selected)
def test_knop_proba():
pool_classifiers, X_dsel, y_dsel, X_test, y_test = setup_classifiers()
knop = KNOP(pool_classifiers)
knop.fit(X_dsel, y_dsel)
probas = knop.predict_proba(X_test)
expected = np.load('deslib/tests/expected_values/knop_proba_integration.npy')
assert np.allclose(probas, expected)
def test_not_predict_proba():
X = X_dsel_ex1
y = y_dsel_ex1
clf1 = Perceptron()
clf1.fit(X, y)
with pytest.raises(ValueError):
knop = KNOP([clf1, clf1])
knop.fit(X, y)
def test_not_predict_proba(create_X_y):
X, y = create_X_y
clf1 = Perceptron()
clf1.fit(X, y)
with pytest.raises(ValueError):
knop = KNOP([clf1, clf1])
knop.fit(X, y)
def test_fit():
knop_test = KNOP(create_pool_classifiers())
knop_test.fit(X_dsel_ex1, y_dsel_ex1)
expected_scores = np.ones(
(15, 6)) * np.array([0.5, 0.5, 1.0, 0.0, 0.33, 0.67])
assert np.array_equal(expected_scores, knop_test.dsel_scores)
# Assert the roc_algorithm is fitted to the scores (decision space) rather than the features (feature space)
assert np.array_equal(knop_test.roc_algorithm._fit_X,
knop_test.dsel_scores)
def test_fit():
knop_test = KNOP(create_pool_classifiers())
knop_test.fit(X_dsel_ex1, y_dsel_ex1)
expected_scores = np.array([[0.5, 0.5], [1.0, 0.0], [0.33, 0.67]])
expected_scores = np.tile(expected_scores, (15, 1, 1))
assert np.array_equal(expected_scores, knop_test.dsel_scores_)
# Assert the roc_algorithm_ is fitted to the scores (decision space) rather than the features (feature space)
expected_roc_data = knop_test.dsel_scores_[:, :, 0]
assert np.array_equal(knop_test.op_knn_._fit_X, expected_roc_data)
def test_fit(example_estimate_competence, create_pool_classifiers):
X, y = example_estimate_competence[0:2]
knop_test = KNOP(create_pool_classifiers)
knop_test.fit(X, y)
expected_scores = np.array([[0.5, 0.5], [1.0, 0.0], [0.33, 0.67]])
expected_scores = np.tile(expected_scores, (15, 1, 1))
assert np.array_equal(expected_scores, knop_test.dsel_scores_)
# Assert the roc_algorithm_ is fitted to the scores (decision space)
# rather than the features (feature space)
expected_roc_data = knop_test.dsel_scores_[:, :, 0]
assert np.array_equal(knop_test.op_knn_._fit_X, expected_roc_data)
def test_estimate_competence(index, expected):
query = np.atleast_2d([1, 1])
knop_test = KNOP(create_pool_classifiers())
knop_test.fit(X_dsel_ex1, y_dsel_ex1)
knop_test.DFP_mask = np.ones(knop_test.n_classifiers)
knop_test.neighbors = neighbors_ex1[index, :]
knop_test.distances = distances_ex1[index, :]
competences = knop_test.estimate_competence(query)
assert np.isclose(competences, expected, atol=0.01).all()
def test_classify(index, expected):
query = np.atleast_2d([1, 1])
knop_test = KNOP(create_pool_classifiers())
knop_test.fit(X_dsel_ex1, y_dsel_ex1)
knop_test.DFP_mask = np.ones(knop_test.n_classifiers)
knop_test.neighbors = neighbors_ex1[index, :]
knop_test.distances = distances_ex1[index, :]
prediction = knop_test.classify_instance(query)
assert prediction == expected
def test_estimate_competence_batch():
query = np.ones((3, 2))
expected = np.array([[4.0, 3.0, 4.0], [5.0, 2.0, 5.0], [2.0, 5.0, 2.0]])
knop_test = KNOP(create_pool_classifiers())
knop_test.fit(X_dsel_ex1, y_dsel_ex1)
knop_test.DFP_mask = np.ones(knop_test.n_classifiers_)
knop_test.neighbors = neighbors_ex1
knop_test._get_similar_out_profiles = Mock(return_value=(None,
neighbors_ex1))
knop_test.distances = distances_ex1
probabilities = np.zeros((3, 6)) # not used in this test
competences = knop_test.estimate_competence_from_proba(
query, probabilities)
assert np.allclose(competences, expected, atol=0.01)
def test_classify(index, expected):
query = np.atleast_2d([1, 1])
knop_test = KNOP(create_pool_classifiers())
knop_test.fit(X_dsel_ex1, y_dsel_ex1)
knop_test.DFP_mask = np.ones(knop_test.n_classifiers)
knop_test.neighbors = neighbors_ex1[index, :]
knop_test.distances = distances_ex1[index, :]
predictions = []
for clf in knop_test.pool_classifiers:
predictions.append(clf.predict(query)[0])
predicted_label = knop_test.classify_instance(query, np.array(predictions))
assert predicted_label == expected
def test_estimate_competence_batch(example_estimate_competence,
create_pool_classifiers):
X, y, neighbors, distances, _, _ = example_estimate_competence
query = np.ones((3, 2))
expected = np.array([[4.0, 3.0, 4.0], [5.0, 2.0, 5.0], [2.0, 5.0, 2.0]])
knop_test = KNOP(create_pool_classifiers)
knop_test.fit(X, y)
knop_test.neighbors = neighbors
knop_test.distances = distances
knop_test._get_similar_out_profiles = Mock(return_value=(None, neighbors))
probabilities = np.zeros((3, 6))
competences = knop_test.estimate_competence_from_proba(
query, probabilities)
assert np.allclose(competences, expected, atol=0.01)
def test_weights_zero():
query = np.atleast_2d([1, 1])
knop_test = KNOP(create_pool_classifiers())
knop_test.fit(X_dsel_ex1, y_dsel_ex1)
knop_test.estimate_competence = MagicMock(return_value=np.zeros(3))
result = knop_test.select(query)
assert np.array_equal(result, np.array([0, 1, 0]))
def test_estimate_competence(index, expected):
query = np.atleast_2d([1, 1])
knop_test = KNOP(create_pool_classifiers())
knop_test.fit(X_dsel_ex1, y_dsel_ex1)
knop_test.DFP_mask = np.ones(knop_test.n_classifiers_)
knop_test.neighbors = neighbors_ex1[index, :]
knop_test._get_similar_out_profiles = Mock(
return_value=(None, np.atleast_2d(neighbors_ex1[index, :])))
knop_test.distances = distances_ex1[index, :]
probabilities = []
for clf in knop_test.pool_classifiers:
probabilities.append(clf.predict_proba(query))
probabilities = np.array(probabilities).transpose((1, 0, 2))
competences = knop_test.estimate_competence_from_proba(
query, probabilities)
assert np.allclose(competences, expected, atol=0.01)
model_perceptron = CalibratedClassifierCV(Perceptron(max_iter=100)).fit(
X_train, y_train)
model_linear_svm = CalibratedClassifierCV(LinearSVC()).fit(
X_train, y_train)
model_svc = SVC(probability=True).fit(X_train, y_train)
model_bayes = GaussianNB().fit(X_train, y_train)
model_tree = DecisionTreeClassifier().fit(X_train, y_train)
model_knn = KNeighborsClassifier(n_neighbors=5).fit(X_train, y_train)
pool_classifiers = [
model_perceptron, model_linear_svm, model_svc, model_bayes, model_tree,
model_knn
]
# Initializing the DS techniques
knop = KNOP(pool_classifiers)
rrc = RRC(pool_classifiers)
lca = LCA(pool_classifiers)
mcb = MCB(pool_classifiers)
aposteriori = APosteriori(pool_classifiers)
# Fitting the techniques
knop.fit(X_dsel, y_dsel)
rrc.fit(X_dsel, y_dsel)
lca.fit(X_dsel, y_dsel)
mcb.fit(X_dsel, y_dsel)
aposteriori.fit(X_dsel, y_dsel)
# Calculate classification accuracy of each technique
print('Evaluating DS techniques:')
print('Classification accuracy KNOP: ', knop.score(X_test, y_test))
def test_check_estimator():
check_estimator(KNOP())
def test_weights_zero():
knop_test = KNOP()
competences = np.zeros((1, 3))
result = knop_test.select(competences)
assert np.all(result)
def test_knop():
pool_classifiers, X_dsel, y_dsel, X_test, y_test = setup_classifiers()
knop = KNOP(pool_classifiers)
knop.fit(X_dsel, y_dsel)
assert np.isclose(knop.score(X_test, y_test), 0.8106060606060606)
def test_knop(knn_methods):
pool_classifiers, X_dsel, y_dsel, X_test, y_test = setup_classifiers()
knop = KNOP(pool_classifiers, knn_classifier=knn_methods)
knop.fit(X_dsel, y_dsel)
assert np.isclose(knop.score(X_test, y_test), 0.97340425531914898)