smile.classification

Class OneVersusOne<T>

java.lang.Object

smile.classification.OneVersusOne<T>

All Implemented Interfaces:

java.io.Serializable, java.util.function.ToDoubleFunction<T>, java.util.function.ToIntFunction<T>, Classifier<T>, SoftClassifier<T>

public class OneVersusOne<T>
extends java.lang.Object
implements SoftClassifier<T>

One-vs-one strategy for reducing the problem of multiclass classification to multiple binary classification problems. This approach trains K (K − 1) / 2 binary classifiers for a K-way multiclass problem; each receives the samples of a pair of classes from the original training set, and must learn to distinguish these two classes. At prediction time, a voting scheme is applied: all K (K − 1) / 2 classifiers are applied to an unseen sample and the class that got the highest number of positive predictions gets predicted by the combined classifier. Like One-vs-rest, one-vs-one suffers from ambiguities in that some regions of its input space may receive the same number of votes.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
OneVersusOne(Classifier<T>[][] classifiers, PlattScaling[][] platts) Constructor.
OneVersusOne(Classifier<T>[][] classifiers, PlattScaling[][] platts, smile.util.IntSet labels) Constructor.

Method Summary

Modifier and Type	Method and Description
static <T> OneVersusOne<T>	fit(T[] x, int[] y, java.util.function.BiFunction<T[],int[],Classifier<T>> trainer) Fits a multi-class model with binary classifiers.
static <T> OneVersusOne<T>	fit(T[] x, int[] y, int pos, int neg, java.util.function.BiFunction<T[],int[],Classifier<T>> trainer) Fits a multi-class model with binary classifiers.
int	predict(T x) Prediction is based on voting.
int	predict(T x, double[] posteriori) Prediction is based posteriori probability estimation.

Methods inherited from class java.lang.Object

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

Methods inherited from interface smile.classification.Classifier

applyAsDouble, applyAsInt, f, predict

Constructor Detail

OneVersusOne

public OneVersusOne(Classifier<T>[][] classifiers,
                    PlattScaling[][] platts)

Constructor.

Parameters:

classifiers - the binary classifier for each one-vs-one case. Only the lower half is needed.

OneVersusOne

public OneVersusOne(Classifier<T>[][] classifiers,
                    PlattScaling[][] platts,
                    smile.util.IntSet labels)

Constructor.

Parameters:

classifiers - the binary classifier for each one-vs-one case. Only the lower half is needed.

labels - the class labels.

Method Detail

fit

public static <T> OneVersusOne<T> fit(T[] x,
                                      int[] y,
                                      java.util.function.BiFunction<T[],int[],Classifier<T>> trainer)

Fits a multi-class model with binary classifiers. Use +1 and -1 as positive and negative class labels.

Parameters:

x - the training samples.

y - the training labels.

trainer - the lambda to train binary classifiers.

fit

public static <T> OneVersusOne<T> fit(T[] x,
                                      int[] y,
                                      int pos,
                                      int neg,
                                      java.util.function.BiFunction<T[],int[],Classifier<T>> trainer)

Fits a multi-class model with binary classifiers.

Parameters:

x - the training samples.

y - the training labels.

pos - the class label for one case.

neg - the class label for rest cases.

trainer - the lambda to train binary classifiers.

predict

public int predict(T x)

Prediction is based on voting.

Specified by:

predict in interface Classifier<T>

Parameters:

x - the instance to be classified.

Returns:

the predicted class label.

predict

public int predict(T x,
                   double[] posteriori)

Prediction is based posteriori probability estimation. The result may be different from predict(T x).

Specified by:

predict in interface SoftClassifier<T>

Parameters:

x - an instance to be classified.

posteriori - the array to store a posteriori probabilities on output.

Returns:

the predicted class label