org.apache.ignite.ml.knn

Class KNNTrainer<M extends KNNModel<Double>,Self extends KNNTrainer<M,Self>>

java.lang.Object

org.apache.ignite.ml.trainers.DatasetTrainer<M,Double>

org.apache.ignite.ml.trainers.SingleLabelDatasetTrainer<M>

org.apache.ignite.ml.knn.KNNTrainer<M,Self>

Type Parameters:

M - Model type.

Direct Known Subclasses:

KNNClassificationTrainer, KNNRegressionTrainer

public abstract class KNNTrainer<M extends KNNModel<Double>,Self extends KNNTrainer<M,Self>>
extends SingleLabelDatasetTrainer<M>

KNN model trader that trains model on top of distribtued spatial indices. Be aware that this model is linked with cluster environment it's been built on and can't be saved or used in other places. Under the hood it keeps Dataset that consists of a set of resources allocated across the cluster.

Nested Class Summary

Nested classes/interfaces inherited from class org.apache.ignite.ml.trainers.DatasetTrainer

DatasetTrainer.EmptyDatasetException

Field Summary

Fields

Modifier and Type	Field and Description
protected DistanceMeasure	distanceMeasure Distance measure.
protected int	k Number of neighbours.
protected boolean	weighted Weighted or not.

Fields inherited from class org.apache.ignite.ml.trainers.DatasetTrainer

envBuilder, environment

Constructor Summary

Constructors

Constructor and Description
KNNTrainer()

Method Summary

Modifier and Type	Method and Description
protected abstract M	convertDatasetIntoModel(Dataset<EmptyContext,SpatialIndex<Double>> dataset) Convers given dataset into KNN model (classification or regression depends on implementation).
protected <K,V> M	fitWithInitializedDeployingContext(DatasetBuilder<K,V> datasetBuilder, Preprocessor<K,V> preprocessor) Trains model based on the specified data.
boolean	isUpdateable(M mdl)
protected abstract Self	self() Returns this instance.
protected <K,V> M	updateModel(M mdl, DatasetBuilder<K,V> datasetBuilder, Preprocessor<K,V> preprocessor) Trains new model taken previous one as a first approximation.
Self	withDataTtl(long dataTtl) Sets up dataTtl parameter.
Self	withDistanceMeasure(DistanceMeasure distanceMeasure) Sets up distanceMeasure parameter.
Self	withIdxType(SpatialIndexType idxType) Sets up idxType parameter.
Self	withK(int k) Sets up k parameter (number of neighbours).
Self	withWeighted(boolean weighted) Sets up weighted parameter.

Methods inherited from class org.apache.ignite.ml.trainers.DatasetTrainer

fit, fit, fit, fit, fit, fit, getLastTrainedModelOrThrowEmptyDatasetException, identityTrainer, learningEnvironment, update, update, update, update, update, withConvertedLabels, withEnvironmentBuilder

Methods inherited from class java.lang.Object

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

Field Detail

distanceMeasure

protected DistanceMeasure distanceMeasure

Distance measure.

k

protected int k

Number of neighbours.

weighted

protected boolean weighted

Weighted or not.

Constructor Detail

KNNTrainer

public KNNTrainer()

Method Detail

fitWithInitializedDeployingContext

protected <K,V> M fitWithInitializedDeployingContext(DatasetBuilder<K,V> datasetBuilder,
                                                     Preprocessor<K,V> preprocessor)

Trains model based on the specified data.

Specified by:

fitWithInitializedDeployingContext in class DatasetTrainer<M extends KNNModel<Double>,Double>

Type Parameters:

K - Type of a key in upstream data.

V - Type of a value in upstream data.

Parameters:

datasetBuilder - Dataset builder.

preprocessor - Extractor of UpstreamEntry into LabeledVector.

Returns:

Model.

isUpdateable

public boolean isUpdateable(M mdl)

Specified by:

isUpdateable in class DatasetTrainer<M extends KNNModel<Double>,Double>

Parameters:

mdl - Model.

Returns:

True if current critical for training parameters correspond to parameters from last training.

updateModel

protected <K,V> M updateModel(M mdl,
                              DatasetBuilder<K,V> datasetBuilder,
                              Preprocessor<K,V> preprocessor)

Trains new model taken previous one as a first approximation.

Specified by:

updateModel in class DatasetTrainer<M extends KNNModel<Double>,Double>

Type Parameters:

K - Type of a key in upstream data.

V - Type of a value in upstream data.

Parameters:

mdl - Learned model.

datasetBuilder - Dataset builder.

preprocessor - Extractor of UpstreamEntry into LabeledVector.

Returns:

Updated model.

convertDatasetIntoModel

protected abstract M convertDatasetIntoModel(Dataset<EmptyContext,SpatialIndex<Double>> dataset)

Convers given dataset into KNN model (classification or regression depends on implementation).

Parameters:

dataset - Dataset of spatial indices.

Returns:

KNN model.

self

protected abstract Self self()

Returns this instance.

Returns:

This instance.

withIdxType

public Self withIdxType(SpatialIndexType idxType)

Sets up idxType parameter.

Parameters:

idxType - Index type.

withDataTtl

public Self withDataTtl(long dataTtl)

Sets up dataTtl parameter.

Parameters:

dataTtl - Partition data time-to-live in seconds (-1 for an infinite lifetime).

Returns:

This instance.

withDistanceMeasure

public Self withDistanceMeasure(DistanceMeasure distanceMeasure)

Sets up distanceMeasure parameter.

Parameters:

distanceMeasure - Distance measure.

withK

public Self withK(int k)

Sets up k parameter (number of neighbours).

Parameters:

k - Number of neighbours.

withWeighted

public Self withWeighted(boolean weighted)

Sets up weighted parameter.

Parameters:

weighted - Weighted or not.