pyanom 0.0.1b1

Anomaly detection package

This library is Python projects for anomaly detection. This contains these techniques.

• Kullback-Leibler desity estimation

• Singular spectrum analysis

• Graphical lasso

• CUMSUM anomaly detection

• Hoteling T2

• Directional data anomaly detection

REQUIREMENTS

• numpy

• pandas

INSTALLATION

pip install pyanom

USAGE

Kullback-Leibler desity estimation

import numpy as np
from pyanom.density_ratio_estimation import KLDensityRatioEstimation

X_normal = np.loadtxt("../input/normal_data.csv", delimiter=",")
X_error = np.loadtxt("../input/error_data.csv", delimiter=",")

model = KLDensityRatioEstimation(
    band_width=0.1, learning_rate=0.1, num_iterations=100)
model.fit(X_normal, X_error)
anomaly_score = model.predict(X_normal, X_error)

Singular spectrum analysis

import numpy as np
from pyanom.subspace_methods import SSA

y_error = np.loadtxt("../input/timeseries_error2.csv", delimiter=",")

model = SSA()
model.fit(y_error, window_size=50, trajectory_n=25, trajectory_pattern=3, test_n=25, test_pattern=2, lag=25)
anomaly_score = model.score()

Graphical lasso

import numpy as np
from pyanom.structure_learning import GraphicalLasso

X_normal = np.loadtxt("../input/normal_data.csv", delimiter=",")
X_error = np.loadtxt("../input/error_data.csv", delimiter=",")

model = GraphicalLasso()
model.fit(X_normal, rho=0.01, normalize=True)
anomaly_score = model.outlier_analysis_score(X_error)

CUSUM anomaly detection

import numpy as np
from pyanom.outlier_detection import CAD

y_normal = np.loadtxt(
    "../input/timeseries_normal.csv", delimiter=",").reshape(-1, 1)
y_error = np.loadtxt(
    "../input/timeseries_error.csv", delimiter=",").reshape(-1, 1)

model = CAD()
model.fit(y_normal, threshold=1)
anomaly_score = model.score(y_error)

Hoteling T2

import numpy as np
from pyanom.outlier_detection import HotelingT2

X_normal = np.loadtxt("../input/normal_data.csv", delimiter=",")
X_error = np.loadtxt("../input/error_data.csv", delimiter=",")

model = HotelingT2()
model.fit(X_normal)
anomaly_score = model.score(X_error)

Directional data anomaly DirectionalDataAnomalyDetection

import numpy as np
from pyanom.outlier_detection import DirectionalDataAnomalyDetection

X_normal = np.loadtxt(
    "../input/normal_direction_data.csv", delimiter=",")
X_error = np.loadtxt("../input/error_direction_data.csv", delimiter=",")

model = DirectionalDataAnomalyDetection()
model.fit(X_normal, normalize=True))
anomaly_score = model.score(X_error)