Package moa.classifiers
Interface OneClassClassifier
-
- All Superinterfaces:
MultiClassClassifier
- All Known Implementing Classes:
Autoencoder
,HSTrees
,NearestNeighbourDescription
public interface OneClassClassifier extends MultiClassClassifier
An interface for incremental classifier models. As an extension of MultiClassClassifier, these classifiers appear in the GUI Classification Tab. Marking them as OneClassClassifiers also allows them to appear specifically when one-class classifiers are needed.- Author:
- Richard Hugh Moulton
-
-
Method Summary
All Methods Instance Methods Abstract Methods Modifier and Type Method Description double
getAnomalyScore(Instance inst)
For use when an anomaly score is needed instead of a vote.void
initialize(Collection<Instance> trainingPoints)
Allows a one class classifier to be initialized with a starting set of training instances.
-
-
-
Method Detail
-
initialize
void initialize(Collection<Instance> trainingPoints)
Allows a one class classifier to be initialized with a starting set of training instances.
-
getAnomalyScore
double getAnomalyScore(Instance inst)
For use when an anomaly score is needed instead of a vote. The higher an instance's anomaly score is, the more likely it is an anomaly.
-
-