org.apache.ignite.ml.knn.classification

Class KNNClassificationModel

java.lang.Object

org.apache.ignite.ml.knn.NNClassificationModel

org.apache.ignite.ml.knn.classification.KNNClassificationModel

All Implemented Interfaces:

Serializable, Function<Vector,Double>, Exportable<KNNModelFormat>, IgniteFunction<Vector,Double>, Model<Vector,Double>

Direct Known Subclasses:

KNNRegressionModel

public class KNNClassificationModel
extends NNClassificationModel
implements Exportable<KNNModelFormat>

kNN algorithm model to solve multi-class classification task.

See Also:

Serialized Form

Field Summary

Fields inherited from class org.apache.ignite.ml.knn.NNClassificationModel

distanceMeasure, k, stgy

Constructor Summary

Constructors

Constructor and Description
KNNClassificationModel(Dataset<EmptyContext,LabeledVectorSet<Double,LabeledVector>> dataset) Builds the model via prepared dataset.

Method Summary

Modifier and Type	Method and Description
Double	apply(Vector v)
void	copyStateFrom(KNNClassificationModel mdl) Copy parameters from other model and save all datasets from it.
protected List<LabeledVector>	findKNearestNeighbors(Vector v) The main idea is calculation all distance pairs between given vector and all vectors in training set, sorting them and finding k vectors with min distance with the given vector.
<P> void	saveModel(Exporter<KNNModelFormat,P> exporter, P path) Save model by the given path.

Methods inherited from class org.apache.ignite.ml.knn.NNClassificationModel

buildLabeledDatasetOnListOfVectors, copyParametersFrom, equals, getClassVoteForVector, getClassWithMaxVotes, getDistanceMeasure, getDistances, getKClosestVectors, hashCode, putDistanceIdxPair, toString, toString, withDistanceMeasure, withK, withStrategy

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface org.apache.ignite.ml.Model

combine

Methods inherited from interface java.util.function.Function

andThen, compose, identity

Constructor Detail

KNNClassificationModel

public KNNClassificationModel(Dataset<EmptyContext,LabeledVectorSet<Double,LabeledVector>> dataset)

Builds the model via prepared dataset.

Parameters:

dataset - Specially prepared object to run algorithm over it.

Method Detail

apply

public Double apply(Vector v)

Specified by:

apply in interface Function<Vector,Double>

saveModel

public <P> void saveModel(Exporter<KNNModelFormat,P> exporter,
                          P path)

Description copied from interface: Exportable

Save model by the given path.

Specified by:

saveModel in interface Exportable<KNNModelFormat>

Specified by:

saveModel in class NNClassificationModel

Parameters:

exporter - Exporter.

path - Path to saved model.

findKNearestNeighbors

protected List<LabeledVector> findKNearestNeighbors(Vector v)

The main idea is calculation all distance pairs between given vector and all vectors in training set, sorting them and finding k vectors with min distance with the given vector.

Parameters:

v - The given vector.

Returns:

K-nearest neighbors.

copyStateFrom

public void copyStateFrom(KNNClassificationModel mdl)

Copy parameters from other model and save all datasets from it.

Parameters:

mdl - Model.