Package org.tribuo.anomaly

package org.tribuo.anomaly

Provides classes and infrastructure for anomaly detection problems.

The anomaly detection Output implementation is called Event and has associated Evaluators, OutputFactory and OutputInfos.

Event trainers are allowed to throw IllegalArgumentException if they are supplied an Event.EventType.ANOMALOUS at training time. It's noted in the documentation if they do support training from anomalous and expected data.

Note: This implementation is effectively a specialised binary classification problem.

Related Packages

Package	Description
org.tribuo	Provides the core interfaces and classes for using Tribuo.
org.tribuo.anomaly.evaluation	Evaluation classes for anomaly detection.
org.tribuo.anomaly.example	Provides anomaly data generators used for demos and testing implementations.
org.tribuo.anomaly.liblinear	Provides an interface to LibLinear-java for anomaly detection problems.
org.tribuo.anomaly.libsvm	Provides an interface to LibSVM for anomaly detection problems.

Class	Description
AnomalyFactory	A factory for generating events.
AnomalyFactory.AnomalyFactoryProvenance	Provenance for AnomalyFactory.
AnomalyInfo	The base class for tracking anomalous events.
Event	An Output representing either an Event.EventType.ANOMALOUS or an Event.EventType.EXPECTED event.
Event.EventType	The type of event.
ImmutableAnomalyInfo	An ImmutableOutputInfo object for Events.
MutableAnomalyInfo	An MutableOutputInfo object for Events.