def test_reference_window(test_path):
from sklearn.utils import shuffle
from pysad.models.integrations import ReferenceWindowModel
from pysad.utils import Data
from pysad.evaluation import AUROCMetric
from pysad.utils import ArrayStreamer
import os
from pyod.models.iforest import IForest
data = Data(os.path.join(test_path,"../../examples/data"))
X_all, y_all = data.get_data("arrhythmia.mat")
X_all, y_all = shuffle(X_all, y_all)
model = ReferenceWindowModel(model_cls=IForest, window_size=240, sliding_size=30,
initial_window_X=X_all[:100])
iterator = ArrayStreamer(shuffle=False)
auroc = AUROCMetric()
y_pred = []
for X, y in iterator.iter(X_all[100:], y_all[100:]):
model.fit_partial(X)
score = model.score_partial(X)
y_pred.append(score)
auroc.update(y, score)
print("AUROC: ", auroc.get())
def test_one_fit(test_path):
from sklearn.utils import shuffle
from pysad.utils import Data
from pysad.evaluation import AUROCMetric
from pysad.utils import ArrayStreamer
import os
from pyod.models.iforest import IForest
from pysad.models.integrations.one_fit_model import OneFitModel
data = Data(os.path.join(test_path, "../../examples/data"))
X_all, y_all = data.get_data("arrhythmia.mat")
print(X_all, y_all)
X_all, y_all = shuffle(X_all, y_all)
model = OneFitModel(model_cls=IForest, initial_X=X_all[:100])
iterator = ArrayStreamer(shuffle=False)
auroc = AUROCMetric()
y_pred = []
for X, y in iterator.iter(X_all[100:], y_all[100:]):
model.fit_partial(X)
score = model.score_partial(X)
y_pred.append(score)
auroc.update(y, score)
print("AUROC: ", auroc.get())
def ensembler_usage_example():
np.random.seed(61) # Fix random seed.
data = Data("data")
X_all, y_all = data.get_data("arrhythmia.mat") # Load Aryhytmia data.
X_all, y_all = shuffle(X_all, y_all) # Shuffle data.
iterator = ArrayStreamer(
shuffle=False) # Create streamer to simulate streaming data.
auroc = AUROCMetric(
) # Tracker of area under receiver-operating- characteristics curve metric.
# Models to be ensembled.
models = [xStream(), LODA()]
ensembler = AverageScoreEnsembler() # Ensembler module.
for X, y in tqdm(iterator.iter(X_all, y_all)): # Iterate over examples.
model_scores = np.empty(len(models), dtype=np.float64)
# Fit & Score via for each model.
for i, model in enumerate(models):
model.fit_partial(X)
model_scores[i] = model.score_partial(X)
score = ensembler.fit_transform_partial(
model_scores) # Fit to ensembler model and get ensembled score.
auroc.update(y, score) # Update AUROC metric.
# Output score.
print("AUROC: {}.".format(auroc.get()))
def full_usage_example():
np.random.seed(61) # Fix random seed.
# Get data to stream.
data = Data("data")
X_all, y_all = data.get_data("arrhythmia.mat")
X_all, y_all = shuffle(X_all, y_all)
iterator = ArrayStreamer(
shuffle=False) # Init streamer to simulate streaming data.
model = xStream() # Init xStream anomaly detection model.
preprocessor = InstanceUnitNormScaler() # Init normalizer.
postprocessor = RunningAveragePostprocessor(
window_size=5) # Init running average postprocessor.
auroc = AUROCMetric(
) # Init area under receiver-operating- characteristics curve metric.
for X, y in tqdm(iterator.iter(X_all[100:], y_all[100:])): # Stream data.
X = preprocessor.fit_transform_partial(
X) # Fit preprocessor to and transform the instance.
score = model.fit_score_partial(
X) # Fit model to and score the instance.
score = postprocessor.fit_transform_partial(
score) # Apply running averaging to the score.
auroc.update(y, score) # Update AUROC metric.
# Output resulting AUROCS metric.
print("AUROC: {}.".format(auroc.get()))
def PyOD_integration_example():
np.random.seed(61) # Fix seed.
# Get data to stream.
data = Data("data")
X_all, y_all = data.get_data("arrhythmia.mat")
X_all, y_all = shuffle(X_all, y_all)
iterator = ArrayStreamer(shuffle=False)
# Fit reference window integration to first 100 instances initially.
model = ReferenceWindowModel(model_cls=IForest,
window_size=240,
sliding_size=30,
initial_window_X=X_all[:100])
auroc = AUROCMetric(
) # Init area under receiver-operating-characteristics curve metric tracker.
for X, y in tqdm(iterator.iter(X_all[100:], y_all[100:])):
model.fit_partial(X) # Fit to the instance.
score = model.score_partial(X) # Score the instance.
auroc.update(y, score) # Update the metric.
# Output AUROC metric.
print("AUROC: {}.".format(auroc.get()))
# Import modules.
from pysad.evaluation import AUROCMetric
from pysad.models import LODA
from pysad.utils import Data
model = LODA() # Init model
metric = AUROCMetric(
) # Init area under receiver-operating- characteristics curve metric
streaming_data = Data().get_iterator("arrhythmia.mat") # Get data streamer.
for x, y_true in streaming_data: # Stream data.
anomaly_score = model.fit_score_partial(
x) # Fit the instance to model and score the instance.
metric.update(y_true, anomaly_score) # Update the AUROC metric.
# Output the resulting AUROCMetric.
print(f"Area under ROC metric is {metric.get()}.")
if __name__ == "__main__":
np.random.seed(61) # Fix random seed.
# Get data to stream.
data = Data("data")
X_all, y_all = data.get_data("arrhythmia.mat")
X_all, y_all = shuffle(X_all, y_all)
iterator = ArrayStreamer(
shuffle=False) # Init streamer to simulate streaming data.
model = xStream() # Init xStream anomaly detection model.
preprocessor = InstanceUnitNormScaler() # Init normalizer.
postprocessor = RunningAveragePostprocessor(
window_size=5) # Init running average postprocessor.
auroc = AUROCMetric(
) # Init area under receiver-operating- characteristics curve metric.
for X, y in tqdm(iterator.iter(X_all[100:], y_all[100:])): # Stream data.
X = preprocessor.fit_transform_partial(
X) # Fit preprocessor to and transform the instance.
score = model.fit_score_partial(
X) # Fit model to and score the instance.
score = postprocessor.fit_transform_partial(
score) # Apply running averaging to the score.
auroc.update(y, score) # Update AUROC metric.
# Output resulting AUROCS metric.
print("AUROC: ", auroc.get())
# iterator = PandasStreamer(shuffle=False)
iterator = ArrayStreamer(shuffle=False)
# model = xStream() # Init xStream anomaly detection model.
# model = IForestASD(initial_window_X=df[:4096], window_size=2048)
model = KitNet()
model.fit(X[:5000])
# need our own preprocessing as hexadecimals dont need zero variance but packet vs checksum length might
preprocessor = InstanceUnitNormScaler() # Init normalizer.
postprocessor = RunningAveragePostprocessor(
window_size=5) # Init running average postprocessor.
# Davies-Bouldin Index, Calinski-Harabasz Index, Silhouette Coefficient exist for clustering, none for anomaly detection
# wrapped in BaseSKLearnMetric
# no metrics I can see, AUC would be perfect if the dataset were labelled
# maybe something else?
auroc = AUROCMetric(
) # Init area under receiver-operating- characteristics curve metric.
for X in tqdm(iterator.iter(X[5000:])): # Stream data.
score = model.score_partial(X)
print(score)
# score = postprocessor.fit_transform_partial(score) # Apply running averaging to the score.
#
# auroc.update(y, score) # Update AUROC metric.
# Output resulting AUROCS metric.
# print("AUROC: ", auroc.get())