def concept_drift_detection(self, X, Y):
if self.init_drift_detection:
if self.drift_detector == "KSWIN":
self.cdd = [KSWIN(w_size = 100, stat_size = 30, alpha=self.confidence) for elem in X.T]
if self.drift_detector == "ADWIN":
self.cdd = [ADWIN() for elem in X.T]
if self.drift_detector == "DDM":
self.cdd = [DDM() for elem in X.T]
if self.drift_detector == "EDDM":
self.cdd = [EDDM() for elem in X.T]
if self.drift_detector == "KSVEC":
self.cdd = KSVEC(vec_size=X.shape[1])
self.init_drift_detection = False
self.drift_detected = False
if not self.init_drift_detection:
if self.drift_detector == "KSVEC":
self.cdd.add_element(X)
if self.cdd.detected_change():
self.drift_detected = True
else:
for elem, detector in zip(X.T, self.cdd):
for e in elem:
detector.add_element(e)
if detector.detected_change():
self.drift_detected = True
self.n_detections = self.n_detections +1
return self.drift_detected
def concept_drift_detection(self, X, Y):
if self.init_drift_detection:
if self.drift_detector == "KSWIN":
self.cdd = [
KSWIN(w_size=100, stat_size=30, alpha=self.confidence)
for elem in X.T
]
if self.drift_detector == "ADWIN":
self.cdd = [ADWIN() for elem in X.T]
if self.drift_detector == "DDM":
self.cdd = [DDM() for elem in X.T]
if self.drift_detector == "EDDM":
self.cdd = [EDDM() for elem in X.T]
self.init_drift_detection = False
self.drift_detected = False
if not self.init_drift_detection:
for elem, detector in zip(X.T, self.cdd):
for e in elem:
detector.add_element(e)
if detector.detected_change():
self.drift_detected = True
self.n_detections = self.n_detections + 1
return self.drift_detected
# if name=="__main__":
# from skmultiflow import
def make_detector(warn=False, s=1e-5, drift_detector='adwin'):
sensitivity = s * 10 if warn else s
if drift_detector == 'adwin':
return ADWIN(delta=sensitivity)
if drift_detector == 'EDDM':
return EDDM()
if drift_detector == 'DDM':
return DDM()
def __init__(self,
list_drifts_detectors = [[ADWIN(delta=0.5), ADWIN(delta=0.05)],
[DDM(min_num_instances=100, warning_level=2.0, out_control_level=3.0)],
[PageHinkley(min_instances=100, delta=0.05, threshold=50, alpha=0.9999), PageHinkley(min_instances=100, delta=0.05, threshold=100, alpha=0.9999)]],
# [SeqDrift2ChangeDetector(delta=0.5, block_size=100), SeqDrift2ChangeDetector(delta=0.05, block_size=100)]],
list_names_drifts_detectors = ['ADWIN','DDM','PH'],
stream = BernoulliStream(drift_period = 1000, n_drifts = 50, widths_drifts = [1], mean_errors = [0.1,0.9]),
n_runs = 1,
name_file = None):
self.name_file = name_file
if (self.name_file == None) or not(isinstance(self.name_file, str)) :
raise ValueError("Attribute 'name_file' must be specified and must be a string, passed {}".format(type(self.name_file)))
self.list_drifts_detectors = list_drifts_detectors
self.list_names_drifts_detectors = list_names_drifts_detectors
self.stream = stream
self.n_runs = n_runs
self.lenght_stream = stream.n_samples
self.true_positions = stream.list_positions
self.n_drift_detectors = len(self.list_drifts_detectors)
def tst():
ddm = DDM()
data_stream = np.random.randint(2, size=2000)
average_prediciton = []
for i in range(1000, 1200):
data_stream[i] = 1
for i in range(1500, 1700):
data_stream[i] = 1
for i in range(2000):
ddm.add_element(data_stream[i])
if ddm.detected_warning_zone():
print('Warning zone has been detected in data: ' +
str(data_stream[i]) + ' - of index: ' + str(i))
if ddm.detected_change():
print('Change has been detected in data: ' + str(data_stream[i]) +
' - of index: ' + str(i))
average_prediciton.append(ddm.miss_prob)
# print('X', hddm.X_, ' ', 'Z', hddm.Z_)
# print('X_e', hddm.X_epsilon, ' ', 'Z', hddm.Z_epsilon)
showPlot(average_prediciton)
[PageHinkley(min_instances=30, delta=0.005, threshold=0.5, alpha=0.999), PageHinkley(min_instances=30, delta=0.005, threshold=1.5, alpha=0.999)], [PageHinkley(min_instances=30, delta=0.005, threshold=0.5, alpha=0.9), PageHinkley(min_instances=30, delta=0.005, threshold=1.5, alpha=0.9)], [PageHinkley(min_instances=30, delta=0.05, threshold=1.5, alpha=0.999), PageHinkley(min_instances=30, delta=0.05, threshold=2.5, alpha=0.999)], [PageHinkley(min_instances=30, delta=0.05, threshold=1.5, alpha=0.9), PageHinkley(min_instances=30, delta=0.05, threshold=2.5, alpha=0.9)], [PageHinkley(min_instances=30, delta=0.05, threshold=0.5, alpha=0.999), PageHinkley(min_instances=30, delta=0.05, threshold=1.5, alpha=0.999)], [PageHinkley(min_instances=30, delta=0.05, threshold=0.5, alpha=0.9), PageHinkley(min_instances=30, delta=0.05, threshold=1.5, alpha=0.9)], [ADWIN(delta=0.5), ADWIN(delta=0.05)], [ADWIN(delta=0.4), ADWIN(delta=0.04)], [ADWIN(delta=0.3), ADWIN(delta=0.03)], [ADWIN(delta=0.2), ADWIN(delta=0.02)], [ADWIN(delta=0.1), ADWIN(delta=0.01)], [ADWIN(delta=0.05), ADWIN(delta=0.005)], [ADWIN(delta=0.02), ADWIN(delta=0.002)], [ADWIN(delta=0.01), ADWIN(delta=0.001)], [ADWIN(delta=0.001), ADWIN(delta=0.0001)], [DDM(min_num_instances=15, warning_level=0.5, out_control_level=1.0)], [DDM(min_num_instances=15, warning_level=0.75, out_control_level=1.25)], [DDM(min_num_instances=15, warning_level=1, out_control_level=1.5)], [DDM(min_num_instances=15, warning_level=1.25, out_control_level=1.75)], [DDM(min_num_instances=15, warning_level=1.5, out_control_level=2.0)], [DDM(min_num_instances=30, warning_level=0.5, out_control_level=1.0)], [DDM(min_num_instances=30, warning_level=0.75, out_control_level=1.25)], [DDM(min_num_instances=30, warning_level=1, out_control_level=1.5)], [DDM(min_num_instances=30, warning_level=1.25, out_control_level=1.75)], [DDM(min_num_instances=30, warning_level=1.5, out_control_level=2.0)], [SeqDrift2ChangeDetector(delta=0.5, block_size=100), SeqDrift2ChangeDetector(delta=0.05, block_size=100)], [SeqDrift2ChangeDetector(delta=0.4, block_size=100), SeqDrift2ChangeDetector(delta=0.04, block_size=100)], [SeqDrift2ChangeDetector(delta=0.3, block_size=100), SeqDrift2ChangeDetector(delta=0.03, block_size=100)], [SeqDrift2ChangeDetector(delta=0.2, block_size=100), SeqDrift2ChangeDetector(delta=0.02, block_size=100)], [SeqDrift2ChangeDetector(delta=0.1, block_size=100), SeqDrift2ChangeDetector(delta=0.01, block_size=100)], [SeqDrift2ChangeDetector(delta=0.05, block_size=100), SeqDrift2ChangeDetector(delta=0.005, block_size=100)],
def __init__(self):
"""Inicjalizacja klasy algorytmu DDM"""
self.name = 'DDM'
self.model = DDM()
self.change_indexes = []
self.warning_zones_indexes = []
######################## CURIE ###################
lst_dim = [n_bins] * n_feats
curie = CA_VonNeumann_Classifier(bins=[],
bins_margin=bins_margin,
dimensions=lst_dim,
cells=empties(lst_dim))
limits_automata = list(np.zeros(1))
#ca_names=['CURIE']
mutants_time = empty_mutant(curie.dimensions)
######################## LEARNERS ###################
learners_ref = [HoeffdingTree(), KNN(), NaiveBayes()]
######################## DETECTORS ###################
detectores_ref = [DDM(), EDDM(), ADWIN(), PageHinkley(), curie]
n_pasos = len(datasets) * len(tipos) * len(learners_ref) * len(
detectores_ref)
SCORES_LER = []
TIMES_LER = []
RAMS_LER = []
DETECTIONS_LER = []
for ler in range(len(learners_ref)):
learner = deepcopy(learners_ref[ler])
SCORES_DET = []
TIMES_DET = []
def reset_model(self):
self.model = DDM()
self.change_indexes = []
self.warning_zones_indexes = []