jsat.classifiers

Class DDAG

java.lang.Object

jsat.classifiers.OneVSOne

jsat.classifiers.DDAG

All Implemented Interfaces:

Serializable, Cloneable, Classifier, Parameterized

public class DDAG
extends OneVSOne

Decision Directed Acyclic Graph (DDAG) classifier. DDAG extends binary decision classifiers into multi-class decision classifiers. Unlike OneVSOne, DDAG results are hard classification decisions, and will not give probabilistic estimates. Accuracy is often very similar, but DDAG classification speed can be significantly faster, as it does not evaluate all possible combinations of classifiers.

Author:

Edward Raff

See Also:

Serialized Form

Field Summary

Fields inherited from class jsat.classifiers.OneVSOne

baseClassifier, oneVone, predicting

Constructor Summary

Constructors

Constructor and Description
DDAG(Classifier baseClassifier) Creates a new DDAG classifier to extend a binary classifier to handle multi-class problems.
DDAG(Classifier baseClassifier, boolean concurrentTrain) Creates a new DDAG classifier to extend a binary classifier to handle multi-class problems.

Method Summary

Modifier and Type	Method and Description
CategoricalResults	classify(DataPoint data) Performs classification on the given data point.
DDAG	clone()

Methods inherited from class jsat.classifiers.OneVSOne

getParameter, getParameters, isConcurrentTraining, setConcurrentTraining, supportsWeightedData, trainC, trainC

Methods inherited from class java.lang.Object

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

Constructor Detail

DDAG

public DDAG(Classifier baseClassifier,
            boolean concurrentTrain)

Creates a new DDAG classifier to extend a binary classifier to handle multi-class problems.

Parameters:

baseClassifier - the binary classifier to extend

concurrentTrain - true to have training of individual classifiers occur in parallel, false to have them use their native parallel training method.

DDAG

public DDAG(Classifier baseClassifier)

Creates a new DDAG classifier to extend a binary classifier to handle multi-class problems.

Parameters:

baseClassifier - the binary classifier to extend

Method Detail

classify

public CategoricalResults classify(DataPoint data)

Description copied from interface: Classifier

Performs classification on the given data point.

Specified by:

classify in interface Classifier

Overrides:

classify in class OneVSOne

Parameters:

data - the data point to classify

Returns:

the results of the classification.

clone

public DDAG clone()

Specified by:

clone in interface Classifier

Overrides:

clone in class OneVSOne