def test_hoeffding_adaptive_tree_categorical_features(test_path):
data_path = os.path.join(test_path, 'ht_categorical_features_testcase.npy')
stream = np.load(data_path)
# Removes the last two columns (regression targets)
stream = stream[:, :-2]
X, y = stream[:, :-1], stream[:, -1]
nominal_attr_idx = np.arange(7).tolist()
learner = HoeffdingAdaptiveTreeClassifier(
nominal_attributes=nominal_attr_idx, random_state=1)
learner.partial_fit(X, y, classes=np.unique(y))
expected_description = "if Attribute 0 = -15.0:\n" \
" Leaf = Class 2 | {2: 475.0}\n" \
"if Attribute 0 = 0.0:\n" \
" Leaf = Class 0 | {0: 560.0, 1: 345.0}\n" \
"if Attribute 0 = 1.0:\n" \
" Leaf = Class 1 | {0: 416.0, 1: 464.0}\n" \
"if Attribute 0 = 2.0:\n" \
" Leaf = Class 1 | {0: 335.0, 1: 504.0}\n" \
"if Attribute 0 = 3.0:\n" \
" Leaf = Class 1 | {0: 244.0, 1: 644.0}\n" \
"if Attribute 0 = -30.0:\n" \
" Leaf = Class 3 | {3: 65.0, 4: 55.0}\n"
assert learner.get_model_description() == expected_description
def test_hoeffding_adaptive_tree_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)
learner = HoeffdingAdaptiveTreeClassifier(leaf_prediction='nb',
random_state=1)
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 = "HoeffdingAdaptiveTreeClassifier(binary_split=False, bootstrap_sampling=True, grace_period=200, " \
"leaf_prediction='nb', max_byte_size=33554432, memory_estimate_period=1000000, nb_threshold=0, " \
"no_preprune=False, nominal_attributes=None, random_state=1, 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
assert type(learner.predict(X)) == np.ndarray
assert type(learner.predict_proba(X)) == np.ndarray
def test_evaluation_event_observer(test_path):
test_file = os.path.join(test_path, 'iris.data')
train_eval_trigger = QuantityBasedHoldoutTrigger(5, 10, 20)
algorithm = HoeffdingAdaptiveTreeClassifier(leaf_prediction='mc',
random_state=1)
results_observer = mock()
evaluation_event_observer = EvaluationEventObserver(
algorithm, train_eval_trigger, [results_observer], [0, 1, 2])
data_source = FileDataSource(record_to_dictionary,
[evaluation_event_observer], test_file)
data_source.listen_for_events()
time.sleep(3)
verify(results_observer, times=4).report(any, any)
def test_hoeffding_adaptive_tree_alternate_tree():
stream = AGRAWALGenerator(random_state=7)
learner = HoeffdingAdaptiveTreeClassifier(random_state=1)
cnt = 0
change_point1 = 1500
change_point2 = 2500
change_point3 = 4000
max_samples = 5000
while cnt < max_samples:
X, y = stream.next_sample()
learner.partial_fit(X, y)
cnt += 1
if cnt > change_point1:
stream.generate_drift()
change_point1 = float('Inf')
expected_description = "if Attribute 2 <= 63.63636363636363:\n" \
" if Attribute 2 <= 39.54545454545455:\n" \
" Leaf = Class 0 | {0: 397.5023676194098}\n" \
" if Attribute 2 > 39.54545454545455:\n" \
" if Attribute 2 <= 58.81818181818181:\n" \
" Leaf = Class 1 | {1: 299.8923824199619}\n" \
" if Attribute 2 > 58.81818181818181:\n" \
" Leaf = Class 0 | {0: 54.0, 1: 20.107617580038095}\n" \
"if Attribute 2 > 63.63636363636363:\n" \
" Leaf = Class 0 | {0: 512.5755895049351}\n"
assert expected_description == learner.get_model_description()
if cnt > change_point2:
stream.generate_drift()
change_point2 = float('Inf')
expected_description = "if Attribute 8 <= 268547.7178694747:\n" \
" Leaf = Class 0 | {0: 446.18690518790413, 1: 80.6180778406834}\n" \
"if Attribute 8 > 268547.7178694747:\n" \
" Leaf = Class 1 | {0: 36.8130948120959, 1: 356.38192215931656}\n"
assert expected_description == learner.get_model_description()
if cnt > change_point3:
stream.generate_drift()
change_point3 = float('Inf')
expected_description = "Leaf = Class 0 | {0: 1083.0, 1: 2.0}\n"
assert expected_description == learner.get_model_description()
total_length = int(total_length)
for data in response.iter_content(chunk_size=4096):
dl += len(data)
f.write(data)
done = int(50 * dl / total_length)
sys.stdout.write("\r[%s%s]" % ('=' * done, ' ' * (50 - done)))
sys.stdout.flush()
data = np.load(file_name, allow_pickle=True)
return data
# data = download_data()
#If dataset file is already downloaded
data = np.load(file_name, allow_pickle=True)
sam = SAMKNN()
arf = HoeffdingAdaptiveTreeClassifier()
stream = DataStream(data[:, 1:], data[:, 0].astype(int))
stream.prepare_for_use()
evaluator = EvaluatePrequential(max_samples=10000,
max_time=1000,
show_plot=True,
metrics=['accuracy', 'kappa'])
evaluator.evaluate(stream=stream,
model=[sam, arf],
model_names=['Sam', 'RSLVQ'])
ds = ConceptDriftStream(random_state=777, position=30000)
ds
# Output:
#ConceptDriftStream(alpha=0.0,
# drift_stream=AGRAWALGenerator(balance_classes=False,
# classification_function=2,
# perturbation=0.0,
# random_state=112),
# position=30000, random_state=777,
# stream=AGRAWALGenerator(balance_classes=False,
# classification_function=0,
# perturbation=0.0, random_state=112),
# width=1000)
# Instantiate the model object
model_hat = HoeffdingAdaptiveTreeClassifier()
# Prequential evaluation
eval1 = EvaluatePrequential(pretrain_size=400,
max_samples=300000,
batch_size=1,
n_wait=100,
max_time=2000,
show_plot=False,
metrics=['accuracy'])
eval1.evaluate(stream=ds, model=model_hat)
# Holdout evaluation
eval2 = EvaluateHoldout(max_samples=30000,
max_time=2000,
from skmultiflow.meta import BatchIncrementalClassifier
all_Files = r'C:\Users\JZM\Desktop\concept-drift-master\concept-drift-master\data\\'
all_Datasets = glob.glob(all_Files + '*.csv')
Results = []
for x in all_Datasets:
print("Results for " + str(os.path.basename(x)))
try:
# Importing Data as stream
data_stream = FileStream(x)
data_stream.prepare_for_use()
# awe = AccuracyWeightedEnsembleClassifier(n_estimators=10,
# base_estimator=HoeffdingTreeClassifier)
# Defining Classifier HT is simple where Adaptive H.Tree uses adwin to minimze error
CLF = [HoeffdingAdaptiveTreeClassifier()]
# Defininf and Training
eval = EvaluatePrequential(show_plot=True,
metrics=[
'accuracy', 'kappa', 'model_size',
'precision', 'recall', 'f1'
],
n_wait=100)
eval.evaluate(stream=data_stream, model=CLF, model_names=['HAT'])
clf = BatchIncrementalClassifier(
base_estimator=HoeffdingTreeClassifier(), n_estimators=10)
# Keeping track of sample count and correct prediction count
sample_count = 0
corrects = 0