def test_predict():
X = X_dsel_ex1
y = y_dsel_ex1
oracle_test = Oracle(create_pool_classifiers())
predicted_labels = oracle_test.predict(X, y)
assert np.equal(predicted_labels, y).all()
assert oracle_test.score(X, y) == 1.0
def test_predict_all_same():
X = X_dsel_ex1
y = np.copy(y_dsel_ex1)
expected = y
oracle_test = Oracle(create_pool_all_agree(return_value=0, size=10))
expected[expected == 1] = -1
predicted_labels = oracle_test.predict(X, y)
assert np.equal(predicted_labels, expected).all()
def test_predict(create_X_y, create_pool_classifiers):
X, y = create_X_y
oracle_test = Oracle(create_pool_classifiers)
oracle_test.fit(X, y)
predicted_labels = oracle_test.predict(X, y)
assert np.equal(predicted_labels, y).all()
assert oracle_test.score(X, y) == 1.0
def test_predict_all_same(create_X_y, create_pool_all_agree):
X, y = create_X_y
expected = y
oracle_test = Oracle(create_pool_all_agree)
oracle_test.fit(X, y)
expected[expected == 1] = 0
predicted_labels = oracle_test.predict(X, y)
assert np.equal(predicted_labels, expected).all()
def test_label_encoder_base_ensemble():
from sklearn.ensemble import RandomForestClassifier
X, y = make_classification()
y[y == 1] = 2
y = y.astype(np.float)
pool = RandomForestClassifier().fit(X, y)
oracle = Oracle(pool)
oracle.fit(X, y)
pred = oracle.predict(X, y)
assert np.isin(oracle.classes_, pred).all()
def test_predict_proba_right_class():
n_test_samples = 200
X, y = make_classification(n_samples=1000)
X_test, y_test = make_classification(n_samples=n_test_samples)
pool = RandomForestClassifier(max_depth=3).fit(X, y)
oracle = Oracle(pool_classifiers=pool).fit(X, y)
preds = oracle.predict(X_test, y_test)
proba = oracle.predict_proba(X_test, y_test)
probas_max = np.argmax(proba, axis=1)
assert np.allclose(probas_max, preds)