ai.h2o.targetencoding

Class TargetEncoderModel

java.lang.Object

water.Iced<T>

water.Keyed<T>

water.Lockable<M>

hex.Model<TargetEncoderModel,TargetEncoderModel.TargetEncoderParameters,TargetEncoderModel.TargetEncoderOutput>

ai.h2o.targetencoding.TargetEncoderModel

All Implemented Interfaces:

java.io.Externalizable, java.io.Serializable, java.lang.Cloneable, water.api.StreamWriter, water.Freezable<TargetEncoderModel>

public class TargetEncoderModel
extends hex.Model<TargetEncoderModel,TargetEncoderModel.TargetEncoderParameters,TargetEncoderModel.TargetEncoderOutput>

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	TargetEncoderModel.DataLeakageHandlingStrategy
static class	TargetEncoderModel.TargetEncoderOutput
static class	TargetEncoderModel.TargetEncoderParameters

Nested classes/interfaces inherited from class hex.Model

hex.Model.AdaptFrameParameters, hex.Model.BigScore, hex.Model.BigScoreChunkPredict, hex.Model.BigScorePredict, hex.Model.Contributions, hex.Model.DeepFeatures, hex.Model.ExemplarMembers, hex.Model.FeatureFrequencies, hex.Model.GetMostImportantFeatures, hex.Model.GetNTrees, hex.Model.GLRMArchetypes, hex.Model.GridSortBy, hex.Model.H2OModelDescriptor, hex.Model.InteractionBuilder, hex.Model.InteractionPair, hex.Model.InteractionSpec, hex.Model.JavaModelStreamWriter, hex.Model.JavaScoringOptions, hex.Model.LeafNodeAssignment, hex.Model.Output, hex.Model.Parameters, hex.Model.PredictScoreResult, hex.Model.RowToTreeAssignment, hex.Model.StagedPredictions, hex.Model.UpdateAuxTreeWeights

Field Summary

Fields

Modifier and Type	Field and Description
static java.lang.String	ALGO_NAME
static int	NO_FOLD

Fields inherited from class hex.Model

_dist, _input_parms, _output, _parms, _toDelete, _warnings, _warningsP, EVAL_AUTO_PARAMS_ENABLED, evalAutoParamsEnabled, scoringInfo

Fields inherited from class water.Lockable

_lockers

Fields inherited from class water.Keyed

_key

Constructor Summary

Constructors

Constructor and Description
TargetEncoderModel(water.Key<TargetEncoderModel> selfKey, TargetEncoderModel.TargetEncoderParameters parms, TargetEncoderModel.TargetEncoderOutput output) Start TargetEncoderModel per se

Method Summary

Modifier and Type	Method and Description
TargetEncoderMojoWriter	getMojo()
hex.ModelMetrics.MetricBuilder	makeMetricBuilder(java.lang.String[] domain)
protected water.Futures	remove_impl(water.Futures fs, boolean cascade)
water.fvec.Frame	score(water.fvec.Frame fr, java.lang.String destination_key, water.Job j, boolean computeMetrics, water.udf.CFuncRef customMetricFunc) Model.score(Frame) always encodes as if the data were new (ie.
protected double[]	score0(double[] data, double[] preds)
water.fvec.Frame	transform(water.fvec.Frame fr)
water.fvec.Frame	transform(water.fvec.Frame fr, BlendingParams blendingParams, double noiseLevel)
water.fvec.Frame	transform(water.fvec.Frame fr, boolean asTraining, int outOfFold, BlendingParams blendingParams, double noiseLevel) Applies target encoding to unseen data during training.
water.fvec.Frame	transformTraining(water.fvec.Frame fr)
water.fvec.Frame	transformTraining(water.fvec.Frame fr, int outOfFold)

Methods inherited from class hex.Model

adaptFrameForScore, adaptTestForJavaScoring, adaptTestForTrain, adaptTestForTrain, adaptTestForTrain, addMetrics, addModelMetrics, addWarning, auc, AUCPR, checksum_impl, classification_error, compareTo, computeDeviances, containsResponse, data, defaultThreshold, defaultThreshold, deleteCrossValidationFoldAssignment, deleteCrossValidationModels, deleteCrossValidationPreds, deviance, deviance, evaluateAutoModelParameters, exportBinaryModel, exportMojo, fetchAll, fillScoringInfo, getDefaultGridSortBy, getGenModelEncoding, getPojoInterfaces, getToEigenVec, haveMojo, havePojo, importBinaryModel, initActualParamValues, isDistributionHuber, isFeatureUsedInPredict, isFeatureUsedInPredict, isSupervised, last_scored, lift_top_group, likelihood, logloss, loss, mae, makeAdaptFrameParameters, makeBigScoreTask, makeInteraction, makeInteractions, makeInteractions, makePojoWriter, makeSchema, makeScoringDomains, makeScoringNames, makeScoringNames, mean_per_class_error, modelDescriptor, mse, needsPostProcess, postProcessPredictions, predictScoreImpl, r2, readAll_impl, resetThreshold, result, rmsle, score, score, score, score, score, score, score, score0, score0, score0, score0PostProcessSupervised, scoreMetrics, scoring_history, scoringDomains, setInputParms, setupBigScorePredict, testJavaScoring, testJavaScoring, testJavaScoring, testJavaScoring, testJavaScoring, toJava, toJava, toJava, toJavaAlgo, toJavaCheckTooBig, toJavaInit, toJavaModelClassName, toJavaPredictBody, toJavaTransform, toJavaUUID, toMojo, toMojo, toString, uploadBinaryModel, writeAll_impl, writeTo

Methods inherited from class water.Lockable

delete_and_lock, delete_and_lock, delete_and_lock, delete, delete, delete, delete, read_lock, read_lock, read_lock, unlock_all, unlock, unlock, unlock, unlock, update, update, update, write_lock_to_read_lock, write_lock, write_lock, write_lock

Methods inherited from class water.Keyed

checksum_impl, checksum, checksum, getKey, readAll, remove_impl, remove_self_key_impl, remove, remove, remove, remove, remove, remove, removeQuietly, writeAll

Methods inherited from class water.Iced

asBytes, clone, copyOver, frozenType, read, readExternal, readJSON, reloadFromBytes, toJsonBytes, toJsonString, write, writeExternal, writeJSON

Methods inherited from class java.lang.Object

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

Field Detail

ALGO_NAME

public static final java.lang.String ALGO_NAME

See Also:

Constant Field Values

NO_FOLD

public static final int NO_FOLD

See Also:

Constant Field Values

Constructor Detail

TargetEncoderModel

public TargetEncoderModel(water.Key<TargetEncoderModel> selfKey,
                          TargetEncoderModel.TargetEncoderParameters parms,
                          TargetEncoderModel.TargetEncoderOutput output)

Start TargetEncoderModel per se

Method Detail

makeMetricBuilder

public hex.ModelMetrics.MetricBuilder makeMetricBuilder(java.lang.String[] domain)

Specified by:

makeMetricBuilder in class hex.Model<TargetEncoderModel,TargetEncoderModel.TargetEncoderParameters,TargetEncoderModel.TargetEncoderOutput>

transformTraining

public water.fvec.Frame transformTraining(water.fvec.Frame fr)

transformTraining

public water.fvec.Frame transformTraining(water.fvec.Frame fr,
                                          int outOfFold)

transform

public water.fvec.Frame transform(water.fvec.Frame fr)

Overrides:

transform in class hex.Model<TargetEncoderModel,TargetEncoderModel.TargetEncoderParameters,TargetEncoderModel.TargetEncoderOutput>

transform

public water.fvec.Frame transform(water.fvec.Frame fr,
                                  BlendingParams blendingParams,
                                  double noiseLevel)

transform

public water.fvec.Frame transform(water.fvec.Frame fr,
                                  boolean asTraining,
                                  int outOfFold,
                                  BlendingParams blendingParams,
                                  double noiseLevel)

Applies target encoding to unseen data during training. This means that DataLeakageHandlingStrategy is enforced to None. In the context of Target Encoding, transform(Frame, BlendingParams, double) should be used to encode new data. Whereas transformTraining(Frame) should be used to encode training data.

Parameters:

fr - Data to transform

asTraining - true iff transforming training data.

outOfFold - if provided (if not = -1), if asTraining=true, and if the model was trained with Kfold strategy, then the frame will be encoded by aggregating encodings from all folds except this one. This is mainly used during cross-validation.

blendingParams - Parameters for blending. If null, blending parameters from models parameters are loaded. If those are not set, DEFAULT_BLENDING_PARAMS from TargetEncoder class are used.

noiseLevel - Level of noise applied (use -1 for default noise level, 0 to disable noise).

Returns:

An instance of Frame with transformed fr, registered in DKV.

score0

protected double[] score0(double[] data,
                          double[] preds)

Specified by:

score0 in class hex.Model<TargetEncoderModel,TargetEncoderModel.TargetEncoderParameters,TargetEncoderModel.TargetEncoderOutput>

score

public water.fvec.Frame score(water.fvec.Frame fr,
                              java.lang.String destination_key,
                              water.Job j,
                              boolean computeMetrics,
                              water.udf.CFuncRef customMetricFunc)
                       throws java.lang.IllegalArgumentException

Model.score(Frame) always encodes as if the data were new (ie. not training data).

Overrides:

score in class hex.Model<TargetEncoderModel,TargetEncoderModel.TargetEncoderParameters,TargetEncoderModel.TargetEncoderOutput>

Throws:

java.lang.IllegalArgumentException

getMojo

public TargetEncoderMojoWriter getMojo()

Overrides:

getMojo in class hex.Model<TargetEncoderModel,TargetEncoderModel.TargetEncoderParameters,TargetEncoderModel.TargetEncoderOutput>

remove_impl

protected water.Futures remove_impl(water.Futures fs,
                                    boolean cascade)

Overrides:

remove_impl in class hex.Model<TargetEncoderModel,TargetEncoderModel.TargetEncoderParameters,TargetEncoderModel.TargetEncoderOutput>