def test_evaluate_delayed_multi_target_classification_coverage(tmpdir):
# A simple coverage test. Tests for metrics are placed in the corresponding test module.
from skmultiflow.data import MultilabelGenerator
from skmultiflow.meta import MultiOutputLearner
max_samples = 1000
# Stream
data = MultilabelGenerator(n_samples=max_samples, random_state=1)
# Get X and y
X, y = data.next_sample(max_samples)
time = generate_random_dates(seed=1, samples=max_samples)
# Setup temporal stream
stream = TemporalDataStream(X, y, time, ordered=True)
# Learner
mol = MultiOutputLearner()
output_file = os.path.join(str(tmpdir), "prequential_delayed_summary.csv")
metrics = ['hamming_score', 'hamming_loss', 'exact_match', 'j_index']
evaluator = EvaluatePrequentialDelayed(max_samples=max_samples,
metrics=metrics,
output_file=output_file)
evaluator.evaluate(stream=stream, model=[mol], model_names=['MOL1'])
def test_label_combination_hoeffding_tree_nb(test_path):
stream = MultilabelGenerator(n_samples=10000,
n_features=15,
n_targets=3,
n_labels=4,
random_state=112)
stream.prepare_for_use()
learner = LabelCombinationHoeffdingTreeClassifier(n_labels=3,
leaf_prediction='nb')
cnt = 0
max_samples = 5000
predictions = []
proba_predictions = []
wait_samples = 100
while cnt < max_samples:
X, y = stream.next_sample()
# Test every n samples
learner.partial_fit(X, y)
if cnt % wait_samples == 0 and (cnt != 0):
predictions.append(learner.predict(X)[0].tolist())
proba_predictions.append(learner.predict_proba(X)[0])
cnt += 1
print(predictions)
expected_predictions = [[0, 0, 1], [1, 1, 1], [0, 1, 1], [0, 1, 1],
[1, 1, 1], [0, 1, 1], [1, 1, 0], [1, 1, 1],
[1, 1, 1], [1, 1, 1], [0, 0, 0], [0, 1, 0],
[1, 0, 0], [1, 0, 1], [1, 1, 1], [1, 1, 1],
[0, 1, 1], [1, 1, 1], [0, 0, 0], [1, 1, 0],
[1, 0, 0], [1, 0, 1], [1, 1, 1], [0, 0, 1],
[1, 0, 1], [1, 1, 1], [1, 0, 0], [1, 1, 1],
[1, 1, 1], [0, 0, 1], [1, 1, 1], [0, 0, 0],
[0, 1, 0], [1, 1, 1], [0, 1, 1], [1, 1, 0],
[0, 0, 1], [0, 0, 0], [1, 1, 1], [1, 1, 1],
[1, 0, 1], [0, 1, 1], [1, 1, 1], [1, 1, 1],
[0, 1, 0], [0, 1, 0], [1, 1, 1], [1, 1, 1],
[1, 1, 1]]
assert np.alltrue(predictions == expected_predictions)
assert type(learner.predict(X)) == np.ndarray
assert type(learner.predict_proba(X)) == np.ndarray
expected_info = "LabelCombinationHoeffdingTreeClassifier(binary_split=False, grace_period=200, " \
"leaf_prediction='nb', max_byte_size=33554432, memory_estimate_period=1000000, n_labels=3, " \
"nb_threshold=0, no_preprune=False, nominal_attributes=None, remove_poor_atts=False, " \
"split_confidence=1e-07, split_criterion='info_gain', stop_mem_management=False, " \
"tie_threshold=0.05)"
info = " ".join([line.strip() for line in learner.get_info().split()])
assert info == expected_info
def test_lc_hoeffding_tree(test_path):
stream = MultilabelGenerator(n_samples=10000,
n_features=15,
n_targets=3,
n_labels=4,
random_state=112)
stream.prepare_for_use()
learner = LCHT(n_labels=3)
cnt = 0
max_samples = 5000
predictions = []
proba_predictions = []
wait_samples = 100
while cnt < max_samples:
X, y = stream.next_sample()
# Test every n samples
learner.partial_fit(X, y)
if cnt % wait_samples == 0 and (cnt != 0):
predictions.append(learner.predict(X)[0].tolist())
proba_predictions.append(learner.predict_proba(X)[0])
cnt += 1
print(predictions)
expected_predictions = [[0, 0, 1], [1, 1, 1], [0, 1, 1], [0, 1, 1],
[1, 1, 1], [0, 1, 1], [1, 1, 0], [1, 1, 1],
[1, 1, 1], [1, 1, 1], [0, 0, 0], [0, 1, 0],
[1, 0, 0], [1, 0, 1], [1, 1, 1], [1, 1, 1],
[0, 1, 1], [1, 1, 1], [0, 0, 0], [1, 1, 0],
[1, 0, 0], [1, 0, 1], [1, 1, 1], [0, 0, 1],
[1, 0, 1], [1, 1, 1], [1, 0, 0], [1, 1, 1],
[1, 1, 1], [0, 0, 1], [1, 1, 1], [0, 0, 0],
[0, 1, 0], [1, 1, 1], [0, 1, 1], [1, 1, 0],
[0, 0, 1], [0, 0, 0], [1, 1, 1], [1, 1, 1],
[1, 0, 1], [0, 1, 1], [1, 1, 1], [1, 1, 1],
[0, 1, 0], [0, 1, 0], [1, 1, 1], [1, 1, 1],
[1, 1, 1]]
assert np.alltrue(predictions == expected_predictions)
assert type(learner.predict(X)) == np.ndarray
assert type(learner.predict_proba(X)) == np.ndarray
def test_evaluate_multi_target_classification_coverage(tmpdir):
# A simple coverage test. Tests for metrics are placed in the corresponding test module.
from skmultiflow.data import MultilabelGenerator
from skmultiflow.meta import MultiOutputLearner
max_samples = 1000
# Stream
stream = MultilabelGenerator(n_samples=max_samples, random_state=1)
stream.prepare_for_use()
# Learner
mol = MultiOutputLearner()
output_file = os.path.join(str(tmpdir), "prequential_summary.csv")
metrics = ['hamming_score', 'hamming_loss', 'exact_match', 'j_index']
evaluator = EvaluatePrequential(max_samples=max_samples,
metrics=metrics,
output_file=output_file)
evaluator.evaluate(stream=stream, model=[mol], model_names=['MOL'])
def test_label_combination_hoeffding_tree_coverage():
# Cover memory management
max_samples = 10000
max_size_kb = 50
stream = MultilabelGenerator(n_samples=10000,
n_features=15,
n_targets=3,
n_labels=4,
random_state=112)
# Unconstrained model has over 62 kB
learner = LabelCombinationHoeffdingTreeClassifier(
n_labels=3,
leaf_prediction='mc',
memory_estimate_period=200,
max_byte_size=max_size_kb * 2**10)
X, y = stream.next_sample(max_samples)
learner.partial_fit(X, y)
assert calculate_object_size(learner, 'kB') <= max_size_kb
def test_classifier_chains():
stream = MultilabelGenerator(random_state=112, n_targets=3, n_samples=5150)
stream.prepare_for_use()
estimator = SGDClassifier(random_state=112, max_iter=10)
learner = ClassifierChain(base_estimator=estimator, random_state=112)
X, y = stream.next_sample(150)
learner.partial_fit(X, y)
cnt = 0
max_samples = 5000
predictions = []
true_labels = []
wait_samples = 100
correct_predictions = 0
while cnt < max_samples:
X, y = stream.next_sample()
# Test every n samples
if (cnt % wait_samples == 0) and (cnt != 0):
predictions.append(learner.predict(X)[0])
true_labels.append(y[0])
if np.array_equal(y[0], predictions[-1]):
correct_predictions += 1
learner.partial_fit(X, y)
cnt += 1
expected_predictions = [[0.0, 0.0, 1.0], [1.0, 0.0, 0.0], [1.0, 0.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 0.0, 1.0], [1.0, 0.0, 0.0],
[1.0, 0.0, 1.0], [1.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[0.0, 0.0, 0.0], [1.0, 0.0, 0.0], [0.0, 1.0, 1.0],
[0.0, 0.0, 1.0], [0.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [1.0, 1.0, 1.0], [1.0, 0.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 0.0, 0.0], [0.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [0.0, 0.0, 1.0], [0.0, 1.0, 1.0],
[0.0, 1.0, 1.0], [1.0, 1.0, 1.0], [0.0, 0.0, 0.0],
[0.0, 1.0, 0.0], [1.0, 1.0, 1.0], [1.0, 1.0, 0.0],
[0.0, 1.0, 0.0], [1.0, 0.0, 1.0], [1.0, 1.0, 1.0],
[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [1.0, 0.0, 0.0],
[1.0, 1.0, 1.0], [0.0, 1.0, 1.0], [0.0, 0.0, 0.0],
[1.0, 0.0, 1.0], [0.0, 0.0, 1.0], [0.0, 0.0, 0.0],
[1.0, 0.0, 1.0], [0.0, 0.0, 1.0], [0.0, 1.0, 0.0],
[0.0, 0.0, 0.0], [1.0, 1.0, 1.0], [0.0, 0.0, 0.0],
[1.0, 1.0, 1.0]]
expected_correct_predictions = 21
assert np.alltrue(np.array_equal(predictions, expected_predictions))
assert correct_predictions == expected_correct_predictions
assert type(learner.predict(X)) == np.ndarray
def test_multi_output_learner_classifier():
stream = MultilabelGenerator(n_samples=5150,
n_features=15,
n_targets=3,
n_labels=4,
random_state=112)
stream.prepare_for_use()
estimator = SGDClassifier(random_state=112,
tol=1e-3,
max_iter=10,
loss='log')
classifier = MultiOutputLearner(base_estimator=estimator)
X, y = stream.next_sample(150)
classifier.partial_fit(X, y)
cnt = 0
max_samples = 5000
predictions = []
true_labels = []
wait_samples = 100
correct_predictions = 0
while cnt < max_samples:
X, y = stream.next_sample()
# Test every n samples
if (cnt % wait_samples == 0) and (cnt != 0):
predictions.append(classifier.predict(X)[0])
true_labels.append(y[0])
if np.array_equal(y[0], predictions[-1]):
correct_predictions += 1
classifier.partial_fit(X, y)
cnt += 1
if StrictVersion(sklearn_version) < StrictVersion("0.21"):
expected_predictions = [[1.0, 1.0, 1.0], [1.0, 0.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 0.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 0.0, 1.0],
[0.0, 1.0, 0.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 0.0, 0.0],
[0.0, 1.0, 1.0], [1.0, 0.0, 0.0],
[1.0, 0.0, 1.0], [0.0, 1.0, 0.0],
[0.0, 0.0, 1.0], [1.0, 0.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 0.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[0.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 0.0, 0.0], [0.0, 1.0, 1.0],
[0.0, 0.0, 1.0], [1.0, 1.0, 0.0],
[1.0, 0.0, 1.0], [1.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 0.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0]]
assert np.alltrue(np.array_equal(predictions, expected_predictions))
expected_correct_predictions = 26
assert correct_predictions == expected_correct_predictions
expected_performance = 0.7755102040816326
performance = hamming_score(true_labels, predictions)
assert np.isclose(performance, expected_performance)
expected_info = "MultiOutputLearner(base_estimator=SGDClassifier(alpha=0.0001, average=False, " \
"class_weight=None,\n" \
" early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,\n" \
" l1_ratio=0.15, learning_rate='optimal', loss='log', max_iter=10,\n" \
" n_iter=None, n_iter_no_change=5, n_jobs=None, penalty='l2',\n" \
" power_t=0.5, random_state=112, shuffle=True, tol=0.001,\n" \
" validation_fraction=0.1, verbose=0, warm_start=False))"
assert classifier.get_info() == expected_info
else:
expected_predictions = [[1.0, 1.0, 1.0], [1.0, 0.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 0.0, 1.0],
[1.0, 0.0, 1.0], [1.0, 0.0, 1.0],
[0.0, 1.0, 0.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 1.0, 0.0],
[1.0, 1.0, 1.0], [1.0, 0.0, 0.0],
[1.0, 0.0, 1.0], [1.0, 1.0, 1.0],
[1.0, 0.0, 1.0], [1.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 0.0], [1.0, 0.0, 1.0],
[0.0, 1.0, 1.0], [1.0, 1.0, 0.0],
[1.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[1.0, 0.0, 0.0], [0.0, 1.0, 1.0],
[0.0, 0.0, 1.0], [1.0, 1.0, 0.0],
[1.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [1.0, 1.0, 1.0],
[1.0, 0.0, 0.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0]]
np.alltrue(np.array_equal(predictions, expected_predictions))
expected_correct_predictions = 23
assert correct_predictions == expected_correct_predictions
expected_performance = 0.7482993197278911
performance = hamming_score(true_labels, predictions)
assert np.isclose(performance, expected_performance)
expected_info = "MultiOutputLearner(base_estimator=SGDClassifier(alpha=0.0001, average=False, " \
"class_weight=None,\n" \
" early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,\n" \
" l1_ratio=0.15, learning_rate='optimal', loss='log', max_iter=10,\n" \
" n_iter_no_change=5, n_jobs=None, penalty='l2', power_t=0.5,\n" \
" random_state=112, shuffle=True, tol=0.001,\n" \
" validation_fraction=0.1, verbose=0, warm_start=False))"
assert classifier.get_info() == expected_info
assert type(classifier.predict(X)) == np.ndarray
assert type(classifier.predict_proba(X)) == np.ndarray
def test_multi_output_learner():
stream = MultilabelGenerator(n_samples=5150,
n_features=15,
n_targets=3,
n_labels=4,
random_state=112)
stream.prepare_for_use()
classifier = MultiOutputLearner(base_estimator=HoeffdingTree())
X, y = stream.next_sample(150)
classifier.partial_fit(X, y)
cnt = 0
max_samples = 5000
predictions = []
true_labels = []
wait_samples = 100
correct_predictions = 0
while cnt < max_samples:
X, y = stream.next_sample()
# Test every n samples
if (cnt % wait_samples == 0) and (cnt != 0):
predictions.append(classifier.predict(X)[0])
true_labels.append(y[0])
if np.array_equal(y[0], predictions[-1]):
correct_predictions += 1
classifier.partial_fit(X, y)
cnt += 1
perf = hamming_score(true_labels, predictions)
expected_predictions = [[1., 1., 1.], [1., 1., 1.], [1., 1., 1.],
[1., 1., 1.], [1., 1., 1.], [0., 1., 1.],
[1., 0., 1.], [1., 0., 1.], [1., 1., 1.],
[0., 0., 1.], [0., 1., 1.], [0., 1., 1.],
[1., 1., 1.], [0., 1., 1.], [1., 1., 0.],
[1., 1., 1.], [0., 1., 1.], [1., 0., 0.],
[1., 0., 1.], [1., 1., 1.], [1., 0., 1.],
[1., 1., 1.], [1., 1., 1.], [1., 1., 1.],
[1., 1., 1.], [1., 1., 1.], [1., 1., 1.],
[1., 0., 0.], [0., 1., 1.], [1., 1., 0.],
[1., 1., 1.], [0., 1., 1.], [1., 1., 1.],
[0., 1., 1.], [1., 0., 1.], [1., 0., 1.],
[0., 0., 1.], [0., 1., 1.], [1., 1., 0.],
[0., 1., 1.], [1., 1., 1.], [1., 1., 1.],
[1., 0., 1.], [1., 1., 1.], [1., 1., 1.],
[1., 0., 1.], [1., 1., 1.], [1., 1., 1.],
[0., 1., 1.]]
expected_correct_predictions = 32
expected_performance = 0.8503401360544217
assert np.alltrue(np.array_equal(predictions, expected_predictions))
assert np.isclose(expected_performance, perf)
assert correct_predictions == expected_correct_predictions
assert type(classifier.predict(X)) == np.ndarray
assert type(classifier.predict_proba(X)) == np.ndarray
def test_classifier_chains():
seed = 112
stream = MultilabelGenerator(random_state=seed, n_targets=3, n_samples=5150)
stream.prepare_for_use()
estimator = SGDClassifier(random_state=seed, tol=1e-3, max_iter=10)
learner = ClassifierChain(base_estimator=estimator, random_state=seed)
X, y = stream.next_sample(150)
learner.partial_fit(X, y)
cnt = 0
max_samples = 5000
predictions = []
true_labels = []
wait_samples = 100
correct_predictions = 0
while cnt < max_samples:
X, y = stream.next_sample()
# Test every n samples
if (cnt % wait_samples == 0) and (cnt != 0):
predictions.append(learner.predict(X)[0])
true_labels.append(y[0])
if np.array_equal(y[0], predictions[-1]):
correct_predictions += 1
learner.partial_fit(X, y)
cnt += 1
if not sklearn_version.startswith("0.21"):
expected_predictions = [[0.0, 0.0, 1.0], [1.0, 0.0, 0.0], [1.0, 0.0, 1.0], [1.0, 1.0, 1.0], [1.0, 0.0, 1.0],
[1.0, 0.0, 0.0], [1.0, 0.0, 1.0], [1.0, 0.0, 1.0], [0.0, 0.0, 1.0], [0.0, 0.0, 0.0],
[1.0, 0.0, 0.0], [0.0, 1.0, 1.0], [0.0, 0.0, 1.0], [0.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [1.0, 1.0, 1.0], [1.0, 0.0, 1.0], [1.0, 1.0, 1.0], [0.0, 0.0, 0.0],
[0.0, 0.0, 1.0], [0.0, 0.0, 1.0], [0.0, 0.0, 1.0], [0.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 0.0, 0.0], [0.0, 1.0, 0.0], [1.0, 1.0, 1.0], [1.0, 1.0, 0.0],
[0.0, 1.0, 0.0], [1.0, 0.0, 1.0], [1.0, 1.0, 1.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0],
[1.0, 0.0, 0.0], [1.0, 1.0, 1.0], [0.0, 1.0, 1.0], [0.0, 0.0, 0.0], [1.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [0.0, 0.0, 0.0], [1.0, 0.0, 1.0], [0.0, 0.0, 1.0], [0.0, 1.0, 0.0],
[0.0, 0.0, 0.0], [1.0, 1.0, 1.0], [0.0, 0.0, 0.0], [1.0, 1.0, 1.0]]
assert np.alltrue(np.array_equal(predictions, expected_predictions))
expected_correct_predictions = 21
assert correct_predictions == expected_correct_predictions
expected_info = "ClassifierChain(base_estimator=SGDClassifier(alpha=0.0001, average=False, class_weight=None,\n" \
" early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,\n" \
" l1_ratio=0.15, learning_rate='optimal', loss='hinge', max_iter=10,\n" \
" n_iter=None, n_iter_no_change=5, n_jobs=None, penalty='l2',\n" \
" power_t=0.5, random_state=112, shuffle=True, tol=0.001,\n" \
" validation_fraction=0.1, verbose=0, warm_start=False),\n" \
" order=None, random_state=112)"
assert learner.get_info() == expected_info
else:
expected_predictions = [[0.0, 0.0, 1.0], [0.0, 0.0, 0.0], [1.0, 0.0, 1.0], [1.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[1.0, 0.0, 0.0], [1.0, 0.0, 1.0], [1.0, 0.0, 1.0], [0.0, 0.0, 1.0], [0.0, 0.0, 0.0],
[1.0, 0.0, 1.0], [0.0, 0.0, 1.0], [0.0, 0.0, 1.0], [0.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [1.0, 0.0, 1.0], [0.0, 0.0, 0.0], [1.0, 0.0, 1.0], [0.0, 0.0, 0.0],
[0.0, 1.0, 1.0], [0.0, 1.0, 1.0], [0.0, 0.0, 1.0], [0.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[0.0, 1.0, 1.0], [0.0, 1.0, 0.0], [0.0, 1.0, 0.0], [1.0, 1.0, 1.0], [0.0, 1.0, 0.0],
[0.0, 1.0, 1.0], [1.0, 0.0, 1.0], [0.0, 1.0, 1.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0],
[1.0, 0.0, 0.0], [1.0, 1.0, 1.0], [0.0, 1.0, 1.0], [0.0, 0.0, 0.0], [1.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 1.0], [0.0, 1.0, 0.0],
[0.0, 0.0, 0.0], [1.0, 1.0, 1.0], [0.0, 0.0, 0.0], [1.0, 1.0, 1.0]]
assert np.alltrue(np.array_equal(predictions, expected_predictions))
expected_correct_predictions = 26
assert correct_predictions == expected_correct_predictions
expected_info = "ClassifierChain(base_estimator=SGDClassifier(alpha=0.0001, average=False, class_weight=None,\n" \
" early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,\n" \
" l1_ratio=0.15, learning_rate='optimal', loss='hinge', max_iter=10,\n" \
" n_iter_no_change=5, n_jobs=None, penalty='l2', power_t=0.5,\n" \
" random_state=112, shuffle=True, tol=0.001,\n" \
" validation_fraction=0.1, verbose=0, warm_start=False),\n" \
" order=None, random_state=112)"
assert learner.get_info() == expected_info
assert type(learner.predict(X)) == np.ndarray
DynamicWeightedMajorityMultiLabel,
OzaBaggingMLClassifier)
from skmultiflow.bayes import NaiveBayes
from skmultiflow.trees import HoeffdingTreeClassifier
from sklearn.linear_model import Perceptron
from skmultiflow.metrics import hamming_score
# from sklearn.linear_model import SGDClassifier
# Setup a data stream
n_features = 10
n_labels = 5
pretrain_size = 150
max_samples = 10000 + pretrain_size
stream = MultilabelGenerator(n_samples=max_samples,
random_state=1,
n_features=n_features,
n_targets=n_labels)
# Binary Relevance Model
br = MultiOutputLearner(Perceptron())
X, y = stream.next_sample(pretrain_size)
br.partial_fit(X, y, classes=stream.target_values)
stream.restart()
# Dynamic weighted classifier model
X, y = stream.next_sample(pretrain_size)
dwm_ml_base_estimator = MultiOutputLearner(Perceptron())
dwm_ml_base_estimator.partial_fit(X, y, classes=stream.target_values)
dwm_ml = DynamicWeightedMajorityMultiLabel(
labels=n_labels, base_estimator=dwm_ml_base_estimator)
# Imports
from skmultiflow.data import MultilabelGenerator
from skmultiflow.meta import MultiOutputLearner
from skmultiflow.trees import LabelCombinationHoeffdingTreeClassifier, HoeffdingTreeClassifier
from skmultiflow.metrics import hamming_score
# Setting up a data stream
stream = MultilabelGenerator(random_state=1, n_samples=200,
n_targets=5, n_features=10)
# Setup Label Combination Hoeffding Tree classifier
lc_ht = LabelCombinationHoeffdingTreeClassifier(n_labels=stream.n_targets)
# Setup variables to control loop and track performance
n_samples = 0
max_samples = 200
true_labels = []
predicts = []
# Train the estimator with the samples provided by the data stream
while n_samples < max_samples and stream.has_more_samples():
X, y = stream.next_sample()
y_pred = lc_ht.predict(X)
lc_ht.partial_fit(X, y, classes=stream.target_values)
predicts.extend(y_pred)
true_labels.extend(y)
n_samples += 1
# Display results
perf = hamming_score(true_labels, predicts)
print('{} samples analyzed.'.format(n_samples))
def test_classifier_chains():
seed = 112
stream = MultilabelGenerator(random_state=seed,
n_targets=3,
n_samples=5150)
estimator = SGDClassifier(random_state=seed, max_iter=10)
learner = ClassifierChain(base_estimator=estimator, random_state=seed)
X, y = stream.next_sample(150)
learner.partial_fit(X, y)
cnt = 0
max_samples = 5000
predictions = []
true_labels = []
wait_samples = 100
correct_predictions = 0
while cnt < max_samples:
X, y = stream.next_sample()
# Test every n samples
if (cnt % wait_samples == 0) and (cnt != 0):
predictions.append(learner.predict(X)[0])
true_labels.append(y[0])
if np.array_equal(y[0], predictions[-1]):
correct_predictions += 1
learner.partial_fit(X, y)
cnt += 1
if not sklearn_version.startswith("0.21"):
expected_predictions = [[0., 0., 1.], [0., 0., 0.], [1., 0., 1.],
[1., 0., 1.], [0., 0., 1.], [1., 0., 0.],
[1., 0., 1.], [1., 0., 1.], [0., 0., 1.],
[0., 0., 0.], [1., 0., 1.], [0., 0., 1.],
[0., 0., 1.], [0., 0., 1.], [0., 0., 1.],
[0., 0., 1.], [1., 0., 1.], [0., 0., 0.],
[1., 0., 1.], [0., 0., 0.], [0., 1., 1.],
[0., 1., 1.], [0., 0., 1.], [0., 1., 1.],
[0., 1., 1.], [0., 1., 1.], [0., 1., 0.],
[0., 1., 0.], [1., 1., 1.], [0., 1., 0.],
[0., 1., 1.], [1., 0., 1.], [0., 1., 1.],
[0., 0., 0.], [0., 0., 0.], [1., 0., 0.],
[1., 1., 1.], [0., 1., 1.], [0., 0., 0.],
[1., 0., 1.], [0., 0., 1.], [0., 0., 0.],
[0., 0., 0.], [0., 0., 1.], [0., 1., 0.],
[0., 0., 0.], [1., 1., 1.], [0., 0., 0.],
[1., 1., 1.]]
assert np.alltrue(np.array_equal(predictions, expected_predictions))
expected_correct_predictions = 26
assert correct_predictions == expected_correct_predictions
expected_info = "ClassifierChain(base_estimator=SGDClassifier(max_iter=10, " \
"random_state=112), order=None, random_state=112)"
info = " ".join([line.strip() for line in learner.get_info().split()])
assert info == expected_info
else:
expected_predictions = [[0.0, 0.0, 1.0], [0.0, 0.0, 0.0],
[1.0, 0.0, 1.0], [1.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [1.0, 0.0, 0.0],
[1.0, 0.0, 1.0], [1.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [0.0, 0.0, 0.0],
[1.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[1.0, 0.0, 1.0], [0.0, 0.0, 0.0],
[1.0, 0.0, 1.0], [0.0, 0.0, 0.0],
[0.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[0.0, 0.0, 1.0], [0.0, 1.0, 1.0],
[0.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[0.0, 1.0, 0.0], [0.0, 1.0, 0.0],
[1.0, 1.0, 1.0], [0.0, 1.0, 0.0],
[0.0, 1.0, 1.0], [1.0, 0.0, 1.0],
[0.0, 1.0, 1.0], [0.0, 0.0, 0.0],
[0.0, 0.0, 0.0], [1.0, 0.0, 0.0],
[1.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[0.0, 0.0, 0.0], [1.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [0.0, 0.0, 0.0],
[0.0, 0.0, 0.0], [0.0, 0.0, 1.0],
[0.0, 1.0, 0.0], [0.0, 0.0, 0.0],
[1.0, 1.0, 1.0], [0.0, 0.0, 0.0],
[1.0, 1.0, 1.0]]
assert np.alltrue(np.array_equal(predictions, expected_predictions))
expected_correct_predictions = 26
assert correct_predictions == expected_correct_predictions
expected_info = "ClassifierChain(base_estimator=SGDClassifier(max_iter=10, " \
"random_state=112), order=None, random_state=112)"
info = " ".join([line.strip() for line in learner.get_info().split()])
assert info == expected_info
assert type(learner.predict(X)) == np.ndarray
def test_multi_output_learner_classifier():
stream = MultilabelGenerator(n_samples=5150,
n_features=15,
n_targets=3,
n_labels=4,
random_state=112)
estimator = SGDClassifier(random_state=112, max_iter=10, loss='log')
classifier = MultiOutputLearner(base_estimator=estimator)
X, y = stream.next_sample(150)
classifier.partial_fit(X, y)
cnt = 0
max_samples = 5000
predictions = []
true_labels = []
wait_samples = 100
correct_predictions = 0
while cnt < max_samples:
X, y = stream.next_sample()
# Test every n samples
if (cnt % wait_samples == 0) and (cnt != 0):
predictions.append(classifier.predict(X)[0])
true_labels.append(y[0])
if np.array_equal(y[0], predictions[-1]):
correct_predictions += 1
classifier.partial_fit(X, y)
cnt += 1
if LooseVersion(sklearn_version) < LooseVersion("0.21"):
expected_predictions = [[1.0, 1.0, 1.0], [1.0, 0.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 0.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 0.0, 1.0],
[0.0, 1.0, 0.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 0.0, 0.0],
[0.0, 1.0, 1.0], [1.0, 0.0, 0.0],
[1.0, 0.0, 1.0], [0.0, 1.0, 0.0],
[0.0, 0.0, 1.0], [1.0, 0.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 0.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[0.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 0.0, 0.0], [0.0, 1.0, 1.0],
[0.0, 0.0, 1.0], [1.0, 1.0, 0.0],
[1.0, 0.0, 1.0], [1.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 0.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0]]
assert np.alltrue(np.array_equal(predictions, expected_predictions))
expected_correct_predictions = 26
assert correct_predictions == expected_correct_predictions
expected_performance = 0.7755102040816326
performance = hamming_score(true_labels, predictions)
assert np.isclose(performance, expected_performance)
expected_info = "MultiOutputLearner(base_estimator=SGDClassifier(loss='log', " \
"random_state=112))"
info = " ".join(
[line.strip() for line in classifier.get_info().split()])
assert info == expected_info
else:
expected_predictions = [[1.0, 1.0, 1.0], [1.0, 0.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 0.0, 1.0],
[1.0, 0.0, 1.0], [1.0, 0.0, 1.0],
[0.0, 1.0, 0.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 1.0, 0.0],
[1.0, 1.0, 1.0], [1.0, 0.0, 0.0],
[1.0, 0.0, 1.0], [1.0, 1.0, 1.0],
[1.0, 0.0, 1.0], [1.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 0.0], [1.0, 0.0, 1.0],
[0.0, 1.0, 1.0], [1.0, 1.0, 0.0],
[1.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [0.0, 1.0, 1.0],
[1.0, 1.0, 1.0], [1.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[1.0, 0.0, 0.0], [0.0, 1.0, 1.0],
[0.0, 0.0, 1.0], [1.0, 1.0, 0.0],
[1.0, 0.0, 1.0], [0.0, 0.0, 1.0],
[0.0, 0.0, 1.0], [1.0, 1.0, 1.0],
[1.0, 0.0, 0.0], [1.0, 1.0, 1.0],
[0.0, 1.0, 1.0]]
np.alltrue(np.array_equal(predictions, expected_predictions))
expected_correct_predictions = 23
assert correct_predictions == expected_correct_predictions
expected_performance = 0.7482993197278911
performance = hamming_score(true_labels, predictions)
assert np.isclose(performance, expected_performance)
expected_info = "MultiOutputLearner(base_estimator=SGDClassifier(loss='log', " \
"max_iter=10, random_state=112))"
info = " ".join(
[line.strip() for line in classifier.get_info().split()])
assert info == expected_info
assert type(classifier.predict(X)) == np.ndarray
assert type(classifier.predict_proba(X)) == np.ndarray