anomaly

Anomaly detection.

Anomaly detection identifies instances that deviate substantially from normal behavior. In data stream learning, the notion of normal behavior can evolve over time, so detectors must adapt to concept drift while flagging outliers in real time.

Modules

datasets

Collection of built in anomaly detection datasets.

Classes

AdaptiveIsolationForest	Adaptive Isolation Forest for anomaly detection.
Autoencoder	Autoencoder anomaly detector
HalfSpaceTrees	Half-Space Trees
IForestASD	iForestASD
Loda	Loda: Lightweight on-line detector of anomalies We implement a streaming version of Loda that updates the histograms after every window of instances.
OnlineIsolationForest	Online Isolation Forest
RSHash	RS-Hash: subspace outlier detection in linear time with randomized hashing.
RobustRandomCutForest	Robust Random Cut Forest.
StreamRHF	StreamRHF anomaly detector
StreamingIsolationForest	Streaming Isolation Forest anomaly detector.