com.bayesserver.analysis

Class InSampleAnomalyDetection

java.lang.Object

com.bayesserver.analysis.InSampleAnomalyDetection

public final class InSampleAnomalyDetection
extends Object

Detects in-sample anomalies in a data set.

Method Summary

Modifier and Type	Method and Description
static InSampleAnomalyDetection	learn(Network network, EvidenceReaderCommandFactory evidenceReaderCommandFactory, InSampleAnomalyDetectionActions actions, InSampleAnomalyDetectionOptions options) Build the in-sample anomaly detector, which can be used to remove anomalous data from a training data set.
InSampleAnomalyDetectionOutput	test(Evidence evidence) Determines whether a record is anomalous.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Method Detail

learn

public static InSampleAnomalyDetection learn(Network network,
                                             EvidenceReaderCommandFactory evidenceReaderCommandFactory,
                                             InSampleAnomalyDetectionActions actions,
                                             InSampleAnomalyDetectionOptions options)

Build the in-sample anomaly detector, which can be used to remove anomalous data from a training data set.

Parameters:

evidenceReaderCommandFactory - An instance which can create evidence reader commands which may be partitioned.

actions - User supplied functions required during the calculation.

network - The network structure to use to for the anomaly detection model. This can be a simple mixture model (cluster model) of a more complicated model.

options - Options affecting how the detector is learned.

test

public InSampleAnomalyDetectionOutput test(Evidence evidence)

Determines whether a record is anomalous. Typically a record is from the in-sample data used to train this instance.

Parameters:

evidence - Evidence to test.

Returns:

An output instance which indicates whether the record is anomalous.