def test_estimate_competence_kuncheva_ex():
query = np.array([1, 1])
mla_test = MLA([create_base_classifier(return_value=1)], k=k_ex_kuncheva)
mla_test.processed_dsel = dsel_processed_kuncheva
mla_test.dsel_scores = dsel_scores_ex_kuncheva
mla_test.DSEL_target = y_dsel_ex_kuncheva_dependent
mla_test.n_classes = n_classes_ex_kuncheva
mla_test.neighbors = neighbors_ex_kuncheva
mla_test.distances = distances_ex_kuncheva
mla_test.DFP_mask = [1]
competences = mla_test.estimate_competence(query.reshape(1, -1))
assert np.isclose(competences, 0.95, atol=0.01)
def test_estimate_competence_diff_target(index):
query = np.array([1, 1])
mla_test = MLA(create_pool_classifiers())
mla_test.processed_dsel = dsel_processed_ex1
mla_test.DSEL_target = np.ones(15, dtype=int) * 3
mla_test.neighbors = neighbors_ex1[index, :]
mla_test.distances = distances_ex1[index, :]
mla_test.DFP_mask = [1, 1, 1]
expected = [0.0, 0.0, 0.0]
competences = mla_test.estimate_competence(query.reshape(1, -1))
assert np.isclose(competences, expected).all()
def test_estimate_competence(index, expected):
query = np.array([1, 1])
mla_test = MLA(create_pool_classifiers())
mla_test.processed_dsel = dsel_processed_ex1
mla_test.dsel_scores = dsel_scores_all_ones
mla_test.DSEL_target = y_dsel_ex1
mla_test.n_classes = 2
mla_test.neighbors = neighbors_ex1[index, :]
mla_test.distances = distances_all_ones[index, :]
mla_test.DFP_mask = [1, 1, 1]
competences = mla_test.estimate_competence(query.reshape(1, -1))
assert np.isclose(competences, expected).all()
def test_estimate_competence_diff_target(index):
query = np.atleast_2d([1, 1])
mla_test = MLA(create_pool_classifiers())
mla_test.processed_dsel = dsel_processed_ex1
mla_test.DSEL_target = np.ones(15, dtype=int) * 3
mla_test.neighbors = neighbors_ex1[index, :]
mla_test.distances = distances_ex1[index, :]
mla_test.DFP_mask = [1, 1, 1]
expected = [0.0, 0.0, 0.0]
predictions = []
for clf in mla_test.pool_classifiers:
predictions.append(clf.predict(query)[0])
competences = mla_test.estimate_competence(
query, predictions=np.array(predictions))
assert np.isclose(competences, expected).all()
def test_estimate_competence(index, expected):
query = np.atleast_2d([1, 1])
mla_test = MLA(create_pool_classifiers())
mla_test.processed_dsel = dsel_processed_ex1
mla_test.dsel_scores = dsel_scores_all_ones
mla_test.DSEL_target = y_dsel_ex1
mla_test.n_classes = 2
mla_test.neighbors = neighbors_ex1[index, :]
mla_test.distances = distances_all_ones[index, :]
mla_test.DFP_mask = [1, 1, 1]
predictions = []
for clf in mla_test.pool_classifiers:
predictions.append(clf.predict(query)[0])
competences = mla_test.estimate_competence(
query, predictions=np.array(predictions))
assert np.isclose(competences, expected).all()
def test_estimate_competence_kuncheva_ex():
query = np.atleast_2d([1, 1])
mla_test = MLA([create_base_classifier(return_value=1)], k=k_ex_kuncheva)
mla_test.processed_dsel = dsel_processed_kuncheva
mla_test.dsel_scores = dsel_scores_ex_kuncheva
mla_test.DSEL_target = y_dsel_ex_kuncheva_dependent
mla_test.n_classes = n_classes_ex_kuncheva
mla_test.neighbors = neighbors_ex_kuncheva
mla_test.distances = distances_ex_kuncheva
mla_test.DFP_mask = [1]
predictions = []
for clf in mla_test.pool_classifiers:
predictions.append(clf.predict(query)[0])
competences = mla_test.estimate_competence(
query, predictions=np.array(predictions))
assert np.isclose(competences, 0.95, atol=0.01)
def test_estimate_competence_batch():
query = np.array([[1, 1], [1, 1], [1, 1]])
expected = np.array([[0.750, 0.666, 0.750], [0.800, 1.000, 0.800],
[1.000, 0.600, 0.500]])
mla_test = MLA(create_pool_classifiers())
mla_test.processed_dsel = dsel_processed_ex1
mla_test.dsel_scores = dsel_scores_all_ones
mla_test.DSEL_target = y_dsel_ex1
mla_test.n_classes = 2
mla_test.neighbors = neighbors_ex1
mla_test.distances = distances_all_ones
mla_test.DFP_mask = np.ones((3, 3))
predictions = []
for clf in mla_test.pool_classifiers:
predictions.append(clf.predict(query)[0])
competences = mla_test.estimate_competence(
query, predictions=np.array(predictions))
assert np.allclose(competences, expected, atol=0.01)