def test_hat_mc(test_path):
stream = ConceptDriftStream(stream=SEAGenerator(random_state=1,
noise_percentage=0.05),
drift_stream=SEAGenerator(
random_state=2,
classification_function=2,
noise_percentage=0.05),
random_state=1,
position=250,
width=10)
stream.prepare_for_use()
learner = HAT(leaf_prediction='mc')
cnt = 0
max_samples = 1000
y_pred = array('i')
y_proba = []
wait_samples = 20
while cnt < max_samples:
X, y = stream.next_sample()
# Test every n samples
if (cnt % wait_samples == 0) and (cnt != 0):
y_pred.append(learner.predict(X)[0])
y_proba.append(learner.predict_proba(X)[0])
learner.partial_fit(X, y)
cnt += 1
expected_predictions = array('i', [
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1
])
assert np.alltrue(y_pred == expected_predictions)
test_file = os.path.join(test_path, 'test_hoeffding_adaptive_tree_mc.npy')
data = np.load(test_file)
assert np.allclose(y_proba, data)
expected_info = "HAT(binary_split=False, grace_period=200, leaf_prediction='mc',\n" \
" max_byte_size=33554432, memory_estimate_period=1000000, nb_threshold=0,\n" \
" no_preprune=False, nominal_attributes=None, remove_poor_atts=False,\n" \
" split_confidence=1e-07, split_criterion='info_gain',\n" \
" stop_mem_management=False, tie_threshold=0.05)"
assert learner.get_info() == expected_info
expected_model_1 = 'Leaf = Class 1.0 | {0.0: 398.0, 1.0: 1000.0}\n'
assert (learner.get_model_description() == expected_model_1)
assert type(learner.predict(X)) == np.ndarray
assert type(learner.predict_proba(X)) == np.ndarray
stream.restart()
X, y = stream.next_sample(5000)
learner = HAT(max_byte_size=30, leaf_prediction='mc', grace_period=10)
learner.partial_fit(X, y)
def test_hat_nb(test_path):
stream = ConceptDriftStream(stream=SEAGenerator(random_state=1,
noise_percentage=0.05),
drift_stream=SEAGenerator(
random_state=2,
classification_function=2,
noise_percentage=0.05),
random_state=1,
position=250,
width=10)
stream.prepare_for_use()
learner = HAT(leaf_prediction='nb')
cnt = 0
max_samples = 1000
y_pred = array('i')
y_proba = []
wait_samples = 20
while cnt < max_samples:
X, y = stream.next_sample()
# Test every n samples
if (cnt % wait_samples == 0) and (cnt != 0):
y_pred.append(learner.predict(X)[0])
y_proba.append(learner.predict_proba(X)[0])
learner.partial_fit(X, y)
cnt += 1
expected_predictions = array('i', [
1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 0,
1, 0, 0, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1,
1
])
assert np.alltrue(y_pred == expected_predictions)
test_file = os.path.join(test_path, 'test_hoeffding_adaptive_tree_nb.npy')
data = np.load(test_file)
assert np.allclose(y_proba, data)
expected_info = 'HAT: max_byte_size: 33554432 - memory_estimate_period: 1000000 - grace_period: 200' \
' - split_criterion: info_gain - split_confidence: 1e-07 - tie_threshold: 0.05' \
' - binary_split: False - stop_mem_management: False - remove_poor_atts: False' \
' - no_pre_prune: False - leaf_prediction: nb - nb_threshold: 0' \
' - nominal_attributes: [] - '
assert learner.get_info() == expected_info
assert type(learner.predict(X)) == np.ndarray
assert type(learner.predict_proba(X)) == np.ndarray
def test_hat_mc(test_path):
stream = ConceptDriftStream(stream=SEAGenerator(random_state=1,
noise_percentage=0.05),
drift_stream=SEAGenerator(
random_state=2,
classification_function=2,
noise_percentage=0.05),
random_state=1,
position=250,
width=10)
stream.prepare_for_use()
learner = HAT(leaf_prediction='mc')
cnt = 0
max_samples = 1000
y_pred = array('i')
y_proba = []
wait_samples = 20
while cnt < max_samples:
X, y = stream.next_sample()
# Test every n samples
if (cnt % wait_samples == 0) and (cnt != 0):
y_pred.append(learner.predict(X)[0])
y_proba.append(learner.predict_proba(X)[0])
learner.partial_fit(X, y)
cnt += 1
expected_predictions = array('i', [
1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 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, 1, 1, 0, 1, 1, 1, 1, 0, 1, 0, 1, 0,
1
])
assert np.alltrue(y_pred == expected_predictions)
test_file = os.path.join(test_path, 'test_hoeffding_adaptive_tree_mc.npy')
data = np.load(test_file)
assert np.allclose(y_proba, data)
expected_info = 'HAT: max_byte_size: 33554432 - memory_estimate_period: 1000000 - grace_period: 200' \
' - split_criterion: info_gain - split_confidence: 1e-07 - tie_threshold: 0.05' \
' - binary_split: False - stop_mem_management: False - remove_poor_atts: False' \
' - no_pre_prune: False - leaf_prediction: mc - nb_threshold: 0' \
' - nominal_attributes: [] - '
assert learner.get_info() == expected_info
expected_model_1 = 'Leaf = Class 1.0 | {0.0: 0.005295278636481529, 1.0: 1.9947047213635185}\n'
expected_model_2 = 'Leaf = Class 1.0 | {0.0: 0.0052952786364815294, 1.0: 1.9947047213635185}\n'
expected_model_3 = 'Leaf = Class 1.0 | {1.0: 1.9947047213635185, 0.0: 0.0052952786364815294}\n'
assert (learner.get_model_description() == expected_model_1) \
or (learner.get_model_description() == expected_model_2) \
or (learner.get_model_description() == expected_model_3)
stream.restart()
X, y = stream.next_sample(5000)
learner = HAT(max_byte_size=30, leaf_prediction='mc', grace_period=10)
learner.partial_fit(X, y)
def test_HAT(test_path):
stream = RandomTreeGenerator(tree_random_state=23, sample_random_state=12, n_classes=4, n_cat_features=2,
n_num_features=5, n_categories_per_cat_feature=5, max_tree_depth=6, min_leaf_depth=3,
fraction_leaves_per_level=0.15)
stream.prepare_for_use()
nominal_attr_idx = [x for x in range(5, stream.n_features)]
learner = HAT(nominal_attributes=nominal_attr_idx)
cnt = 0
max_samples = 5000
predictions = array('d')
proba_predictions = []
wait_samples = 100
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])
proba_predictions.append(learner.predict_proba(X)[0])
learner.partial_fit(X, y)
cnt += 1
expected_predictions = array('d', [2.0, 1.0, 1.0, 1.0, 0.0, 3.0, 0.0, 1.0, 1.0, 2.0,
0.0, 2.0, 1.0, 1.0, 2.0, 1.0, 3.0, 0.0, 1.0, 1.0,
1.0, 1.0, 0.0, 3.0, 1.0, 2.0, 1.0, 1.0, 3.0, 2.0,
1.0, 2.0, 2.0, 2.0, 1.0, 1.0, 1.0, 0.0, 1.0, 2.0,
0.0, 2.0, 0.0, 0.0, 0.0, 0.0, 1.0, 3.0, 2.0])
test_file = os.path.join(test_path, 'test_hoeffding_adaptive_tree.npy')
data = np.load(test_file)
assert np.alltrue(predictions == expected_predictions)
assert np.allclose(proba_predictions, data)
expected_info = 'HAT: max_byte_size: 33554432 - memory_estimate_period: 1000000 - grace_period: 200' \
' - split_criterion: info_gain - split_confidence: 1e-07 - tie_threshold: 0.05' \
' - binary_split: False - stop_mem_management: False - remove_poor_atts: False' \
' - no_pre_prune: False - leaf_prediction: nba - nb_threshold: 0' \
' - nominal_attributes: [5, 6, 7, 8, 9, 10, 11, 12, 13, 14] - '
assert learner.get_info() == expected_info
expected_model_1 = 'Leaf = Class 1.0 | {0.0: 1367.3628584299263, 1.0: 1702.2738590243584,' \
' 2.0: 952.1668539501372, 3.0: 822.1964285955778}\n'
expected_model_2 = 'Leaf = Class 1.0 | {1.0: 1702.2738590243584, 2.0: 952.1668539501372,' \
' 0.0: 1367.3628584299263, 3.0: 822.1964285955778}\n'
expected_model_3 = 'Leaf = Class 1.0 | {1.0: 1702.2738590243584, 2.0: 952.16685395013724, ' \
'0.0: 1367.3628584299263, 3.0: 822.1964285955778}\n' # Python 3.6
expected_model_4 = 'Leaf = Class 1.0 | {0.0: 1367.3628584299263, 1.0: 1702.2738590243584,' \
' 2.0: 952.16685395013724, 3.0: 822.1964285955778}\n' # Python 3.4
assert (learner.get_model_description() == expected_model_1) \
or (learner.get_model_description() == expected_model_2) \
or (learner.get_model_description() == expected_model_3) \
or (learner.get_model_description() == expected_model_4)
def test_hat_nba(test_path):
stream = HyperplaneGenerator(mag_change=0.001,
noise_percentage=0.1,
random_state=2)
stream.prepare_for_use()
learner = HAT(leaf_prediction='nba')
cnt = 0
max_samples = 5000
y_pred = array('i')
y_proba = []
wait_samples = 100
while cnt < max_samples:
X, y = stream.next_sample()
# Test every n samples
if (cnt % wait_samples == 0) and (cnt != 0):
y_pred.append(learner.predict(X)[0])
y_proba.append(learner.predict_proba(X)[0])
learner.partial_fit(X, y)
cnt += 1
expected_predictions = array('i', [
1, 0, 1, 0, 1, 1, 0, 1, 1, 1, 0, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1,
0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1, 0, 1, 1, 1,
0
])
assert np.alltrue(y_pred == expected_predictions)
test_file = os.path.join(test_path, 'test_hoeffding_adaptive_tree_nba.npy')
data = np.load(test_file)
assert np.allclose(y_proba, data)
expected_info = "HAT(binary_split=False, bootstrap_sampling=True, grace_period=200,\n" \
" leaf_prediction='nba', max_byte_size=33554432,\n" \
" memory_estimate_period=1000000, nb_threshold=0, no_preprune=False,\n" \
" nominal_attributes=None, remove_poor_atts=False, split_confidence=1e-07,\n" \
" split_criterion='info_gain', stop_mem_management=False, tie_threshold=0.05)"
assert learner.get_info() == expected_info
assert type(learner.predict(X)) == np.ndarray
assert type(learner.predict_proba(X)) == np.ndarray
# ### Entrenar el clasificador (modelo)
# In[2]:
clasificador = HAT()
print(clasificador.get_info())
print("start training")
clasificador.fit(X_train, y_train, classes=None, sample_weight=None)
print("end training")
# In[3]:
print("start predict")
predict = clasificador.predict(X_test)
print("end predict")
# In[4]:
print("shape_predict")
print(predict.shape)
print("score")
print(clasificador.score(X_test, y_test))
import matplotlib.pyplot as plt
#%matplotlib inline
plt.rcParams['figure.figsize'] = (16, 9)
plt.style.use('ggplot')
plt.hist([predict, y_test])
plt.show()
