• Home
• com.google.apis
• google-api-services-bigquery
• v2-rev20220422-1.32.1
• source code
• TrainingOptions.html

×

Project download

Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance.
Project price only 1 $

You can buy this project and download/modify it how often you want.

target.apidocs.com.google.api.services.bigquery.model.TrainingOptions.html Maven / Gradle / Ivy

Skip navigation links




Overview
Package
Class
Use
Tree
Deprecated
Index
Help




Prev Class
Next Class


Frames
No Frames


All Classes






Summary: 
Nested | 
Field | 
Constr | 
Method


Detail: 
Field | 
Constr | 
Method








com.google.api.services.bigquery.model
Class TrainingOptions



java.lang.Object


java.util.AbstractMap<String,Object>


com.google.api.client.util.GenericData


com.google.api.client.json.GenericJson


com.google.api.services.bigquery.model.TrainingOptions













All Implemented Interfaces:
Cloneable, Map<String,Object>




public final class TrainingOptions
extends GenericJson
Options used in model training.

 
 This is the Java data model class that specifies how to parse/serialize into the JSON that is
 transmitted over HTTP when working with the BigQuery API. For a detailed explanation see:
 https://developers.google.com/api-client-library/java/google-http-java-client/json
 

Author:
Google, Inc.












Nested Class Summary




Nested classes/interfaces inherited from class com.google.api.client.util.GenericData
GenericData.Flags





Nested classes/interfaces inherited from class java.util.AbstractMap
AbstractMap.SimpleEntry<K,V>, AbstractMap.SimpleImmutableEntry<K,V>





Nested classes/interfaces inherited from interface java.util.Map
Map.Entry<K,V>








Constructor Summary

Constructors 

Constructor and Description


TrainingOptions() 









Method Summary

All Methods Instance Methods Concrete Methods 

Modifier and Type
Method and Description


TrainingOptions
clone() 


Boolean
getAdjustStepChanges()
If true, detect step changes and make data adjustment in the input time series.



Boolean
getAutoArima()
Whether to enable auto ARIMA or not.



Long
getAutoArimaMaxOrder()
The max value of non-seasonal p and q.



Long
getBatchSize()
Batch size for dnn models.



String
getBoosterType()
Booster type for boosted tree models.



Boolean
getCalculatePValues()
Whether or not p-value test should be computed for this model.



Boolean
getCleanSpikesAndDips()
If true, clean spikes and dips in the input time series.



Double
getColsampleBylevel()
Subsample ratio of columns for each level for boosted tree models.



Double
getColsampleBynode()
Subsample ratio of columns for each node(split) for boosted tree models.



Double
getColsampleBytree()
Subsample ratio of columns when constructing each tree for boosted tree models.



String
getDartNormalizeType()
Type of normalization algorithm for boosted tree models using dart booster.



String
getDataFrequency()
The data frequency of a time series.



String
getDataSplitColumn()
The column to split data with.



Double
getDataSplitEvalFraction()
The fraction of evaluation data over the whole input data.



String
getDataSplitMethod()
The data split type for training and evaluation, e.g.



Boolean
getDecomposeTimeSeries()
If true, perform decompose time series and save the results.



String
getDistanceType()
Distance type for clustering models.



Double
getDropout()
Dropout probability for dnn models.



Boolean
getEarlyStop()
Whether to stop early when the loss doesn't improve significantly any more (compared to
 min_relative_progress).



Boolean
getEnableGlobalExplain()
If true, enable global explanation during training.



String
getFeedbackType()
Feedback type that specifies which algorithm to run for matrix factorization.



List<Long>
getHiddenUnits()
Hidden units for dnn models.



String
getHolidayRegion()
The geographical region based on which the holidays are considered in time series modeling.



Long
getHorizon()
The number of periods ahead that need to be forecasted.



List<String>
getHparamTuningObjectives()
The target evaluation metrics to optimize the hyperparameters for.



Boolean
getIncludeDrift()
Include drift when fitting an ARIMA model.



Double
getInitialLearnRate()
Specifies the initial learning rate for the line search learn rate strategy.



List<String>
getInputLabelColumns()
Name of input label columns in training data.



Long
getIntegratedGradientsNumSteps()
Number of integral steps for the integrated gradients explain method.



String
getItemColumn()
Item column specified for matrix factorization models.



String
getKmeansInitializationColumn()
The column used to provide the initial centroids for kmeans algorithm when
 kmeans_initialization_method is CUSTOM.



String
getKmeansInitializationMethod()
The method used to initialize the centroids for kmeans algorithm.



Double
getL1Regularization()
L1 regularization coefficient.



Double
getL2Regularization()
L2 regularization coefficient.



Map<String,Double>
getLabelClassWeights()
Weights associated with each label class, for rebalancing the training data.



Double
getLearnRate()
Learning rate in training.



String
getLearnRateStrategy()
The strategy to determine learn rate for the current iteration.



String
getLossType()
Type of loss function used during training run.



Long
getMaxIterations()
The maximum number of iterations in training.



Long
getMaxParallelTrials()
Maximum number of trials to run in parallel.



Long
getMaxTreeDepth()
Maximum depth of a tree for boosted tree models.



Double
getMinRelativeProgress()
When early_stop is true, stops training when accuracy improvement is less than
 'min_relative_progress'.



Double
getMinSplitLoss()
Minimum split loss for boosted tree models.



Long
getMinTreeChildWeight()
Minimum sum of instance weight needed in a child for boosted tree models.



String
getModelUri()
Google Cloud Storage URI from which the model was imported.



ArimaOrder
getNonSeasonalOrder()
A specification of the non-seasonal part of the ARIMA model: the three components (p, d, q) are
 the AR order, the degree of differencing, and the MA order.



Long
getNumClusters()
Number of clusters for clustering models.



Long
getNumFactors()
Num factors specified for matrix factorization models.



Long
getNumParallelTree()
Number of parallel trees constructed during each iteration for boosted tree models.



Long
getNumTrials()
Number of trials to run this hyperparameter tuning job.



String
getOptimizationStrategy()
Optimization strategy for training linear regression models.



Boolean
getPreserveInputStructs()
Whether to preserve the input structs in output feature names.



Long
getSampledShapleyNumPaths()
Number of paths for the sampled shapley explain method.



Double
getSubsample()
Subsample fraction of the training data to grow tree to prevent overfitting for boosted tree
 models.



String
getTimeSeriesDataColumn()
Column to be designated as time series data for ARIMA model.



String
getTimeSeriesIdColumn()
The time series id column that was used during ARIMA model training.



List<String>
getTimeSeriesIdColumns()
The time series id columns that were used during ARIMA model training.



String
getTimeSeriesTimestampColumn()
Column to be designated as time series timestamp for ARIMA model.



String
getTreeMethod()
Tree construction algorithm for boosted tree models.



String
getUserColumn()
User column specified for matrix factorization models.



Double
getWalsAlpha()
Hyperparameter for matrix factoration when implicit feedback type is specified.



Boolean
getWarmStart()
Whether to train a model from the last checkpoint.



TrainingOptions
set(String fieldName,
   Object value) 


TrainingOptions
setAdjustStepChanges(Boolean adjustStepChanges)
If true, detect step changes and make data adjustment in the input time series.



TrainingOptions
setAutoArima(Boolean autoArima)
Whether to enable auto ARIMA or not.



TrainingOptions
setAutoArimaMaxOrder(Long autoArimaMaxOrder)
The max value of non-seasonal p and q.



TrainingOptions
setBatchSize(Long batchSize)
Batch size for dnn models.



TrainingOptions
setBoosterType(String boosterType)
Booster type for boosted tree models.



TrainingOptions
setCalculatePValues(Boolean calculatePValues)
Whether or not p-value test should be computed for this model.



TrainingOptions
setCleanSpikesAndDips(Boolean cleanSpikesAndDips)
If true, clean spikes and dips in the input time series.



TrainingOptions
setColsampleBylevel(Double colsampleBylevel)
Subsample ratio of columns for each level for boosted tree models.



TrainingOptions
setColsampleBynode(Double colsampleBynode)
Subsample ratio of columns for each node(split) for boosted tree models.



TrainingOptions
setColsampleBytree(Double colsampleBytree)
Subsample ratio of columns when constructing each tree for boosted tree models.



TrainingOptions
setDartNormalizeType(String dartNormalizeType)
Type of normalization algorithm for boosted tree models using dart booster.



TrainingOptions
setDataFrequency(String dataFrequency)
The data frequency of a time series.



TrainingOptions
setDataSplitColumn(String dataSplitColumn)
The column to split data with.



TrainingOptions
setDataSplitEvalFraction(Double dataSplitEvalFraction)
The fraction of evaluation data over the whole input data.



TrainingOptions
setDataSplitMethod(String dataSplitMethod)
The data split type for training and evaluation, e.g.



TrainingOptions
setDecomposeTimeSeries(Boolean decomposeTimeSeries)
If true, perform decompose time series and save the results.



TrainingOptions
setDistanceType(String distanceType)
Distance type for clustering models.



TrainingOptions
setDropout(Double dropout)
Dropout probability for dnn models.



TrainingOptions
setEarlyStop(Boolean earlyStop)
Whether to stop early when the loss doesn't improve significantly any more (compared to
 min_relative_progress).



TrainingOptions
setEnableGlobalExplain(Boolean enableGlobalExplain)
If true, enable global explanation during training.



TrainingOptions
setFeedbackType(String feedbackType)
Feedback type that specifies which algorithm to run for matrix factorization.



TrainingOptions
setHiddenUnits(List<Long> hiddenUnits)
Hidden units for dnn models.



TrainingOptions
setHolidayRegion(String holidayRegion)
The geographical region based on which the holidays are considered in time series modeling.



TrainingOptions
setHorizon(Long horizon)
The number of periods ahead that need to be forecasted.



TrainingOptions
setHparamTuningObjectives(List<String> hparamTuningObjectives)
The target evaluation metrics to optimize the hyperparameters for.



TrainingOptions
setIncludeDrift(Boolean includeDrift)
Include drift when fitting an ARIMA model.



TrainingOptions
setInitialLearnRate(Double initialLearnRate)
Specifies the initial learning rate for the line search learn rate strategy.



TrainingOptions
setInputLabelColumns(List<String> inputLabelColumns)
Name of input label columns in training data.



TrainingOptions
setIntegratedGradientsNumSteps(Long integratedGradientsNumSteps)
Number of integral steps for the integrated gradients explain method.



TrainingOptions
setItemColumn(String itemColumn)
Item column specified for matrix factorization models.



TrainingOptions
setKmeansInitializationColumn(String kmeansInitializationColumn)
The column used to provide the initial centroids for kmeans algorithm when
 kmeans_initialization_method is CUSTOM.



TrainingOptions
setKmeansInitializationMethod(String kmeansInitializationMethod)
The method used to initialize the centroids for kmeans algorithm.



TrainingOptions
setL1Regularization(Double l1Regularization)
L1 regularization coefficient.



TrainingOptions
setL2Regularization(Double l2Regularization)
L2 regularization coefficient.



TrainingOptions
setLabelClassWeights(Map<String,Double> labelClassWeights)
Weights associated with each label class, for rebalancing the training data.



TrainingOptions
setLearnRate(Double learnRate)
Learning rate in training.



TrainingOptions
setLearnRateStrategy(String learnRateStrategy)
The strategy to determine learn rate for the current iteration.



TrainingOptions
setLossType(String lossType)
Type of loss function used during training run.



TrainingOptions
setMaxIterations(Long maxIterations)
The maximum number of iterations in training.



TrainingOptions
setMaxParallelTrials(Long maxParallelTrials)
Maximum number of trials to run in parallel.



TrainingOptions
setMaxTreeDepth(Long maxTreeDepth)
Maximum depth of a tree for boosted tree models.



TrainingOptions
setMinRelativeProgress(Double minRelativeProgress)
When early_stop is true, stops training when accuracy improvement is less than
 'min_relative_progress'.



TrainingOptions
setMinSplitLoss(Double minSplitLoss)
Minimum split loss for boosted tree models.



TrainingOptions
setMinTreeChildWeight(Long minTreeChildWeight)
Minimum sum of instance weight needed in a child for boosted tree models.



TrainingOptions
setModelUri(String modelUri)
Google Cloud Storage URI from which the model was imported.



TrainingOptions
setNonSeasonalOrder(ArimaOrder nonSeasonalOrder)
A specification of the non-seasonal part of the ARIMA model: the three components (p, d, q) are
 the AR order, the degree of differencing, and the MA order.



TrainingOptions
setNumClusters(Long numClusters)
Number of clusters for clustering models.



TrainingOptions
setNumFactors(Long numFactors)
Num factors specified for matrix factorization models.



TrainingOptions
setNumParallelTree(Long numParallelTree)
Number of parallel trees constructed during each iteration for boosted tree models.



TrainingOptions
setNumTrials(Long numTrials)
Number of trials to run this hyperparameter tuning job.



TrainingOptions
setOptimizationStrategy(String optimizationStrategy)
Optimization strategy for training linear regression models.



TrainingOptions
setPreserveInputStructs(Boolean preserveInputStructs)
Whether to preserve the input structs in output feature names.



TrainingOptions
setSampledShapleyNumPaths(Long sampledShapleyNumPaths)
Number of paths for the sampled shapley explain method.



TrainingOptions
setSubsample(Double subsample)
Subsample fraction of the training data to grow tree to prevent overfitting for boosted tree
 models.



TrainingOptions
setTimeSeriesDataColumn(String timeSeriesDataColumn)
Column to be designated as time series data for ARIMA model.



TrainingOptions
setTimeSeriesIdColumn(String timeSeriesIdColumn)
The time series id column that was used during ARIMA model training.



TrainingOptions
setTimeSeriesIdColumns(List<String> timeSeriesIdColumns)
The time series id columns that were used during ARIMA model training.



TrainingOptions
setTimeSeriesTimestampColumn(String timeSeriesTimestampColumn)
Column to be designated as time series timestamp for ARIMA model.



TrainingOptions
setTreeMethod(String treeMethod)
Tree construction algorithm for boosted tree models.



TrainingOptions
setUserColumn(String userColumn)
User column specified for matrix factorization models.



TrainingOptions
setWalsAlpha(Double walsAlpha)
Hyperparameter for matrix factoration when implicit feedback type is specified.



TrainingOptions
setWarmStart(Boolean warmStart)
Whether to train a model from the last checkpoint.







Methods inherited from class com.google.api.client.json.GenericJson
getFactory, setFactory, toPrettyString, toString





Methods inherited from class com.google.api.client.util.GenericData
entrySet, equals, get, getClassInfo, getUnknownKeys, hashCode, put, putAll, remove, setUnknownKeys





Methods inherited from class java.util.AbstractMap
clear, containsKey, containsValue, isEmpty, keySet, size, values





Methods inherited from class java.lang.Object
finalize, getClass, notify, notifyAll, wait, wait, wait





Methods inherited from interface java.util.Map
compute, computeIfAbsent, computeIfPresent, forEach, getOrDefault, merge, putIfAbsent, remove, replace, replace, replaceAll














Constructor Detail





TrainingOptions
public TrainingOptions()









Method Detail





getAdjustStepChanges
public Boolean getAdjustStepChanges()
If true, detect step changes and make data adjustment in the input time series.

Returns:
value or null for none








setAdjustStepChanges
public TrainingOptions setAdjustStepChanges(Boolean adjustStepChanges)
If true, detect step changes and make data adjustment in the input time series.

Parameters:
adjustStepChanges - adjustStepChanges or null for none








getAutoArima
public Boolean getAutoArima()
Whether to enable auto ARIMA or not.

Returns:
value or null for none








setAutoArima
public TrainingOptions setAutoArima(Boolean autoArima)
Whether to enable auto ARIMA or not.

Parameters:
autoArima - autoArima or null for none








getAutoArimaMaxOrder
public Long getAutoArimaMaxOrder()
The max value of non-seasonal p and q.

Returns:
value or null for none








setAutoArimaMaxOrder
public TrainingOptions setAutoArimaMaxOrder(Long autoArimaMaxOrder)
The max value of non-seasonal p and q.

Parameters:
autoArimaMaxOrder - autoArimaMaxOrder or null for none








getBatchSize
public Long getBatchSize()
Batch size for dnn models.

Returns:
value or null for none








setBatchSize
public TrainingOptions setBatchSize(Long batchSize)
Batch size for dnn models.

Parameters:
batchSize - batchSize or null for none








getBoosterType
public String getBoosterType()
Booster type for boosted tree models.

Returns:
value or null for none








setBoosterType
public TrainingOptions setBoosterType(String boosterType)
Booster type for boosted tree models.

Parameters:
boosterType - boosterType or null for none








getCalculatePValues
public Boolean getCalculatePValues()
Whether or not p-value test should be computed for this model. Only available for linear and
 logistic regression models.

Returns:
value or null for none








setCalculatePValues
public TrainingOptions setCalculatePValues(Boolean calculatePValues)
Whether or not p-value test should be computed for this model. Only available for linear and
 logistic regression models.

Parameters:
calculatePValues - calculatePValues or null for none








getCleanSpikesAndDips
public Boolean getCleanSpikesAndDips()
If true, clean spikes and dips in the input time series.

Returns:
value or null for none








setCleanSpikesAndDips
public TrainingOptions setCleanSpikesAndDips(Boolean cleanSpikesAndDips)
If true, clean spikes and dips in the input time series.

Parameters:
cleanSpikesAndDips - cleanSpikesAndDips or null for none








getColsampleBylevel
public Double getColsampleBylevel()
Subsample ratio of columns for each level for boosted tree models.

Returns:
value or null for none








setColsampleBylevel
public TrainingOptions setColsampleBylevel(Double colsampleBylevel)
Subsample ratio of columns for each level for boosted tree models.

Parameters:
colsampleBylevel - colsampleBylevel or null for none








getColsampleBynode
public Double getColsampleBynode()
Subsample ratio of columns for each node(split) for boosted tree models.

Returns:
value or null for none








setColsampleBynode
public TrainingOptions setColsampleBynode(Double colsampleBynode)
Subsample ratio of columns for each node(split) for boosted tree models.

Parameters:
colsampleBynode - colsampleBynode or null for none








getColsampleBytree
public Double getColsampleBytree()
Subsample ratio of columns when constructing each tree for boosted tree models.

Returns:
value or null for none








setColsampleBytree
public TrainingOptions setColsampleBytree(Double colsampleBytree)
Subsample ratio of columns when constructing each tree for boosted tree models.

Parameters:
colsampleBytree - colsampleBytree or null for none








getDartNormalizeType
public String getDartNormalizeType()
Type of normalization algorithm for boosted tree models using dart booster.

Returns:
value or null for none








setDartNormalizeType
public TrainingOptions setDartNormalizeType(String dartNormalizeType)
Type of normalization algorithm for boosted tree models using dart booster.

Parameters:
dartNormalizeType - dartNormalizeType or null for none








getDataFrequency
public String getDataFrequency()
The data frequency of a time series.

Returns:
value or null for none








setDataFrequency
public TrainingOptions setDataFrequency(String dataFrequency)
The data frequency of a time series.

Parameters:
dataFrequency - dataFrequency or null for none








getDataSplitColumn
public String getDataSplitColumn()
The column to split data with. This column won't be used as a feature. 1. When
 data_split_method is CUSTOM, the corresponding column should be boolean. The rows with true
 value tag are eval data, and the false are training data. 2. When data_split_method is SEQ, the
 first DATA_SPLIT_EVAL_FRACTION rows (from smallest to largest) in the corresponding column are
 used as training data, and the rest are eval data. It respects the order in Orderable data
 types: https://cloud.google.com/bigquery/docs/reference/standard-sql/data-types#data-type-
 properties

Returns:
value or null for none








setDataSplitColumn
public TrainingOptions setDataSplitColumn(String dataSplitColumn)
The column to split data with. This column won't be used as a feature. 1. When
 data_split_method is CUSTOM, the corresponding column should be boolean. The rows with true
 value tag are eval data, and the false are training data. 2. When data_split_method is SEQ, the
 first DATA_SPLIT_EVAL_FRACTION rows (from smallest to largest) in the corresponding column are
 used as training data, and the rest are eval data. It respects the order in Orderable data
 types: https://cloud.google.com/bigquery/docs/reference/standard-sql/data-types#data-type-
 properties

Parameters:
dataSplitColumn - dataSplitColumn or null for none








getDataSplitEvalFraction
public Double getDataSplitEvalFraction()
The fraction of evaluation data over the whole input data. The rest of data will be used as
 training data. The format should be double. Accurate to two decimal places. Default value is
 0.2.

Returns:
value or null for none








setDataSplitEvalFraction
public TrainingOptions setDataSplitEvalFraction(Double dataSplitEvalFraction)
The fraction of evaluation data over the whole input data. The rest of data will be used as
 training data. The format should be double. Accurate to two decimal places. Default value is
 0.2.

Parameters:
dataSplitEvalFraction - dataSplitEvalFraction or null for none








getDataSplitMethod
public String getDataSplitMethod()
The data split type for training and evaluation, e.g. RANDOM.

Returns:
value or null for none








setDataSplitMethod
public TrainingOptions setDataSplitMethod(String dataSplitMethod)
The data split type for training and evaluation, e.g. RANDOM.

Parameters:
dataSplitMethod - dataSplitMethod or null for none








getDecomposeTimeSeries
public Boolean getDecomposeTimeSeries()
If true, perform decompose time series and save the results.

Returns:
value or null for none








setDecomposeTimeSeries
public TrainingOptions setDecomposeTimeSeries(Boolean decomposeTimeSeries)
If true, perform decompose time series and save the results.

Parameters:
decomposeTimeSeries - decomposeTimeSeries or null for none








getDistanceType
public String getDistanceType()
Distance type for clustering models.

Returns:
value or null for none








setDistanceType
public TrainingOptions setDistanceType(String distanceType)
Distance type for clustering models.

Parameters:
distanceType - distanceType or null for none








getDropout
public Double getDropout()
Dropout probability for dnn models.

Returns:
value or null for none








setDropout
public TrainingOptions setDropout(Double dropout)
Dropout probability for dnn models.

Parameters:
dropout - dropout or null for none








getEarlyStop
public Boolean getEarlyStop()
Whether to stop early when the loss doesn't improve significantly any more (compared to
 min_relative_progress). Used only for iterative training algorithms.

Returns:
value or null for none








setEarlyStop
public TrainingOptions setEarlyStop(Boolean earlyStop)
Whether to stop early when the loss doesn't improve significantly any more (compared to
 min_relative_progress). Used only for iterative training algorithms.

Parameters:
earlyStop - earlyStop or null for none








getEnableGlobalExplain
public Boolean getEnableGlobalExplain()
If true, enable global explanation during training.

Returns:
value or null for none








setEnableGlobalExplain
public TrainingOptions setEnableGlobalExplain(Boolean enableGlobalExplain)
If true, enable global explanation during training.

Parameters:
enableGlobalExplain - enableGlobalExplain or null for none








getFeedbackType
public String getFeedbackType()
Feedback type that specifies which algorithm to run for matrix factorization.

Returns:
value or null for none








setFeedbackType
public TrainingOptions setFeedbackType(String feedbackType)
Feedback type that specifies which algorithm to run for matrix factorization.

Parameters:
feedbackType - feedbackType or null for none








getHiddenUnits
public List<Long> getHiddenUnits()
Hidden units for dnn models.

Returns:
value or null for none








setHiddenUnits
public TrainingOptions setHiddenUnits(List<Long> hiddenUnits)
Hidden units for dnn models.

Parameters:
hiddenUnits - hiddenUnits or null for none








getHolidayRegion
public String getHolidayRegion()
The geographical region based on which the holidays are considered in time series modeling. If
 a valid value is specified, then holiday effects modeling is enabled.

Returns:
value or null for none








setHolidayRegion
public TrainingOptions setHolidayRegion(String holidayRegion)
The geographical region based on which the holidays are considered in time series modeling. If
 a valid value is specified, then holiday effects modeling is enabled.

Parameters:
holidayRegion - holidayRegion or null for none








getHorizon
public Long getHorizon()
The number of periods ahead that need to be forecasted.

Returns:
value or null for none








setHorizon
public TrainingOptions setHorizon(Long horizon)
The number of periods ahead that need to be forecasted.

Parameters:
horizon - horizon or null for none








getHparamTuningObjectives
public List<String> getHparamTuningObjectives()
The target evaluation metrics to optimize the hyperparameters for.

Returns:
value or null for none








setHparamTuningObjectives
public TrainingOptions setHparamTuningObjectives(List<String> hparamTuningObjectives)
The target evaluation metrics to optimize the hyperparameters for.

Parameters:
hparamTuningObjectives - hparamTuningObjectives or null for none








getIncludeDrift
public Boolean getIncludeDrift()
Include drift when fitting an ARIMA model.

Returns:
value or null for none








setIncludeDrift
public TrainingOptions setIncludeDrift(Boolean includeDrift)
Include drift when fitting an ARIMA model.

Parameters:
includeDrift - includeDrift or null for none








getInitialLearnRate
public Double getInitialLearnRate()
Specifies the initial learning rate for the line search learn rate strategy.

Returns:
value or null for none








setInitialLearnRate
public TrainingOptions setInitialLearnRate(Double initialLearnRate)
Specifies the initial learning rate for the line search learn rate strategy.

Parameters:
initialLearnRate - initialLearnRate or null for none








getInputLabelColumns
public List<String> getInputLabelColumns()
Name of input label columns in training data.

Returns:
value or null for none








setInputLabelColumns
public TrainingOptions setInputLabelColumns(List<String> inputLabelColumns)
Name of input label columns in training data.

Parameters:
inputLabelColumns - inputLabelColumns or null for none








getIntegratedGradientsNumSteps
public Long getIntegratedGradientsNumSteps()
Number of integral steps for the integrated gradients explain method.

Returns:
value or null for none








setIntegratedGradientsNumSteps
public TrainingOptions setIntegratedGradientsNumSteps(Long integratedGradientsNumSteps)
Number of integral steps for the integrated gradients explain method.

Parameters:
integratedGradientsNumSteps - integratedGradientsNumSteps or null for none








getItemColumn
public String getItemColumn()
Item column specified for matrix factorization models.

Returns:
value or null for none








setItemColumn
public TrainingOptions setItemColumn(String itemColumn)
Item column specified for matrix factorization models.

Parameters:
itemColumn - itemColumn or null for none








getKmeansInitializationColumn
public String getKmeansInitializationColumn()
The column used to provide the initial centroids for kmeans algorithm when
 kmeans_initialization_method is CUSTOM.

Returns:
value or null for none








setKmeansInitializationColumn
public TrainingOptions setKmeansInitializationColumn(String kmeansInitializationColumn)
The column used to provide the initial centroids for kmeans algorithm when
 kmeans_initialization_method is CUSTOM.

Parameters:
kmeansInitializationColumn - kmeansInitializationColumn or null for none








getKmeansInitializationMethod
public String getKmeansInitializationMethod()
The method used to initialize the centroids for kmeans algorithm.

Returns:
value or null for none








setKmeansInitializationMethod
public TrainingOptions setKmeansInitializationMethod(String kmeansInitializationMethod)
The method used to initialize the centroids for kmeans algorithm.

Parameters:
kmeansInitializationMethod - kmeansInitializationMethod or null for none








getL1Regularization
public Double getL1Regularization()
L1 regularization coefficient.

Returns:
value or null for none








setL1Regularization
public TrainingOptions setL1Regularization(Double l1Regularization)
L1 regularization coefficient.

Parameters:
l1Regularization - l1Regularization or null for none








getL2Regularization
public Double getL2Regularization()
L2 regularization coefficient.

Returns:
value or null for none








setL2Regularization
public TrainingOptions setL2Regularization(Double l2Regularization)
L2 regularization coefficient.

Parameters:
l2Regularization - l2Regularization or null for none








getLabelClassWeights
public Map<String,Double> getLabelClassWeights()
Weights associated with each label class, for rebalancing the training data. Only applicable
 for classification models.

Returns:
value or null for none








setLabelClassWeights
public TrainingOptions setLabelClassWeights(Map<String,Double> labelClassWeights)
Weights associated with each label class, for rebalancing the training data. Only applicable
 for classification models.

Parameters:
labelClassWeights - labelClassWeights or null for none








getLearnRate
public Double getLearnRate()
Learning rate in training. Used only for iterative training algorithms.

Returns:
value or null for none








setLearnRate
public TrainingOptions setLearnRate(Double learnRate)
Learning rate in training. Used only for iterative training algorithms.

Parameters:
learnRate - learnRate or null for none








getLearnRateStrategy
public String getLearnRateStrategy()
The strategy to determine learn rate for the current iteration.

Returns:
value or null for none








setLearnRateStrategy
public TrainingOptions setLearnRateStrategy(String learnRateStrategy)
The strategy to determine learn rate for the current iteration.

Parameters:
learnRateStrategy - learnRateStrategy or null for none








getLossType
public String getLossType()
Type of loss function used during training run.

Returns:
value or null for none








setLossType
public TrainingOptions setLossType(String lossType)
Type of loss function used during training run.

Parameters:
lossType - lossType or null for none








getMaxIterations
public Long getMaxIterations()
The maximum number of iterations in training. Used only for iterative training algorithms.

Returns:
value or null for none








setMaxIterations
public TrainingOptions setMaxIterations(Long maxIterations)
The maximum number of iterations in training. Used only for iterative training algorithms.

Parameters:
maxIterations - maxIterations or null for none








getMaxParallelTrials
public Long getMaxParallelTrials()
Maximum number of trials to run in parallel.

Returns:
value or null for none








setMaxParallelTrials
public TrainingOptions setMaxParallelTrials(Long maxParallelTrials)
Maximum number of trials to run in parallel.

Parameters:
maxParallelTrials - maxParallelTrials or null for none








getMaxTreeDepth
public Long getMaxTreeDepth()
Maximum depth of a tree for boosted tree models.

Returns:
value or null for none








setMaxTreeDepth
public TrainingOptions setMaxTreeDepth(Long maxTreeDepth)
Maximum depth of a tree for boosted tree models.

Parameters:
maxTreeDepth - maxTreeDepth or null for none








getMinRelativeProgress
public Double getMinRelativeProgress()
When early_stop is true, stops training when accuracy improvement is less than
 'min_relative_progress'. Used only for iterative training algorithms.

Returns:
value or null for none








setMinRelativeProgress
public TrainingOptions setMinRelativeProgress(Double minRelativeProgress)
When early_stop is true, stops training when accuracy improvement is less than
 'min_relative_progress'. Used only for iterative training algorithms.

Parameters:
minRelativeProgress - minRelativeProgress or null for none








getMinSplitLoss
public Double getMinSplitLoss()
Minimum split loss for boosted tree models.

Returns:
value or null for none








setMinSplitLoss
public TrainingOptions setMinSplitLoss(Double minSplitLoss)
Minimum split loss for boosted tree models.

Parameters:
minSplitLoss - minSplitLoss or null for none








getMinTreeChildWeight
public Long getMinTreeChildWeight()
Minimum sum of instance weight needed in a child for boosted tree models.

Returns:
value or null for none








setMinTreeChildWeight
public TrainingOptions setMinTreeChildWeight(Long minTreeChildWeight)
Minimum sum of instance weight needed in a child for boosted tree models.

Parameters:
minTreeChildWeight - minTreeChildWeight or null for none








getModelUri
public String getModelUri()
Google Cloud Storage URI from which the model was imported. Only applicable for imported
 models.

Returns:
value or null for none








setModelUri
public TrainingOptions setModelUri(String modelUri)
Google Cloud Storage URI from which the model was imported. Only applicable for imported
 models.

Parameters:
modelUri - modelUri or null for none








getNonSeasonalOrder
public ArimaOrder getNonSeasonalOrder()
A specification of the non-seasonal part of the ARIMA model: the three components (p, d, q) are
 the AR order, the degree of differencing, and the MA order.

Returns:
value or null for none








setNonSeasonalOrder
public TrainingOptions setNonSeasonalOrder(ArimaOrder nonSeasonalOrder)
A specification of the non-seasonal part of the ARIMA model: the three components (p, d, q) are
 the AR order, the degree of differencing, and the MA order.

Parameters:
nonSeasonalOrder - nonSeasonalOrder or null for none








getNumClusters
public Long getNumClusters()
Number of clusters for clustering models.

Returns:
value or null for none








setNumClusters
public TrainingOptions setNumClusters(Long numClusters)
Number of clusters for clustering models.

Parameters:
numClusters - numClusters or null for none








getNumFactors
public Long getNumFactors()
Num factors specified for matrix factorization models.

Returns:
value or null for none








setNumFactors
public TrainingOptions setNumFactors(Long numFactors)
Num factors specified for matrix factorization models.

Parameters:
numFactors - numFactors or null for none








getNumParallelTree
public Long getNumParallelTree()
Number of parallel trees constructed during each iteration for boosted tree models.

Returns:
value or null for none








setNumParallelTree
public TrainingOptions setNumParallelTree(Long numParallelTree)
Number of parallel trees constructed during each iteration for boosted tree models.

Parameters:
numParallelTree - numParallelTree or null for none








getNumTrials
public Long getNumTrials()
Number of trials to run this hyperparameter tuning job.

Returns:
value or null for none








setNumTrials
public TrainingOptions setNumTrials(Long numTrials)
Number of trials to run this hyperparameter tuning job.

Parameters:
numTrials - numTrials or null for none








getOptimizationStrategy
public String getOptimizationStrategy()
Optimization strategy for training linear regression models.

Returns:
value or null for none








setOptimizationStrategy
public TrainingOptions setOptimizationStrategy(String optimizationStrategy)
Optimization strategy for training linear regression models.

Parameters:
optimizationStrategy - optimizationStrategy or null for none








getPreserveInputStructs
public Boolean getPreserveInputStructs()
Whether to preserve the input structs in output feature names. Suppose there is a struct A with
 field b. When false (default), the output feature name is A_b. When true, the output feature
 name is A.b.

Returns:
value or null for none








setPreserveInputStructs
public TrainingOptions setPreserveInputStructs(Boolean preserveInputStructs)
Whether to preserve the input structs in output feature names. Suppose there is a struct A with
 field b. When false (default), the output feature name is A_b. When true, the output feature
 name is A.b.

Parameters:
preserveInputStructs - preserveInputStructs or null for none








getSampledShapleyNumPaths
public Long getSampledShapleyNumPaths()
Number of paths for the sampled shapley explain method.

Returns:
value or null for none








setSampledShapleyNumPaths
public TrainingOptions setSampledShapleyNumPaths(Long sampledShapleyNumPaths)
Number of paths for the sampled shapley explain method.

Parameters:
sampledShapleyNumPaths - sampledShapleyNumPaths or null for none








getSubsample
public Double getSubsample()
Subsample fraction of the training data to grow tree to prevent overfitting for boosted tree
 models.

Returns:
value or null for none








setSubsample
public TrainingOptions setSubsample(Double subsample)
Subsample fraction of the training data to grow tree to prevent overfitting for boosted tree
 models.

Parameters:
subsample - subsample or null for none








getTimeSeriesDataColumn
public String getTimeSeriesDataColumn()
Column to be designated as time series data for ARIMA model.

Returns:
value or null for none








setTimeSeriesDataColumn
public TrainingOptions setTimeSeriesDataColumn(String timeSeriesDataColumn)
Column to be designated as time series data for ARIMA model.

Parameters:
timeSeriesDataColumn - timeSeriesDataColumn or null for none








getTimeSeriesIdColumn
public String getTimeSeriesIdColumn()
The time series id column that was used during ARIMA model training.

Returns:
value or null for none








setTimeSeriesIdColumn
public TrainingOptions setTimeSeriesIdColumn(String timeSeriesIdColumn)
The time series id column that was used during ARIMA model training.

Parameters:
timeSeriesIdColumn - timeSeriesIdColumn or null for none








getTimeSeriesIdColumns
public List<String> getTimeSeriesIdColumns()
The time series id columns that were used during ARIMA model training.

Returns:
value or null for none








setTimeSeriesIdColumns
public TrainingOptions setTimeSeriesIdColumns(List<String> timeSeriesIdColumns)
The time series id columns that were used during ARIMA model training.

Parameters:
timeSeriesIdColumns - timeSeriesIdColumns or null for none








getTimeSeriesTimestampColumn
public String getTimeSeriesTimestampColumn()
Column to be designated as time series timestamp for ARIMA model.

Returns:
value or null for none








setTimeSeriesTimestampColumn
public TrainingOptions setTimeSeriesTimestampColumn(String timeSeriesTimestampColumn)
Column to be designated as time series timestamp for ARIMA model.

Parameters:
timeSeriesTimestampColumn - timeSeriesTimestampColumn or null for none








getTreeMethod
public String getTreeMethod()
Tree construction algorithm for boosted tree models.

Returns:
value or null for none








setTreeMethod
public TrainingOptions setTreeMethod(String treeMethod)
Tree construction algorithm for boosted tree models.

Parameters:
treeMethod - treeMethod or null for none








getUserColumn
public String getUserColumn()
User column specified for matrix factorization models.

Returns:
value or null for none








setUserColumn
public TrainingOptions setUserColumn(String userColumn)
User column specified for matrix factorization models.

Parameters:
userColumn - userColumn or null for none








getWalsAlpha
public Double getWalsAlpha()
Hyperparameter for matrix factoration when implicit feedback type is specified.

Returns:
value or null for none








setWalsAlpha
public TrainingOptions setWalsAlpha(Double walsAlpha)
Hyperparameter for matrix factoration when implicit feedback type is specified.

Parameters:
walsAlpha - walsAlpha or null for none








getWarmStart
public Boolean getWarmStart()
Whether to train a model from the last checkpoint.

Returns:
value or null for none








setWarmStart
public TrainingOptions setWarmStart(Boolean warmStart)
Whether to train a model from the last checkpoint.

Parameters:
warmStart - warmStart or null for none








set
public TrainingOptions set(String fieldName,
                           Object value)

Overrides:
set in class GenericJson








clone
public TrainingOptions clone()

Overrides:
clone in class GenericJson














Skip navigation links




Overview
Package
Class
Use
Tree
Deprecated
Index
Help




Prev Class
Next Class


Frames
No Frames


All Classes






Summary: 
Nested | 
Field | 
Constr | 
Method


Detail: 
Field | 
Constr | 
Method






Copyright © 2011–2022 Google. All rights reserved.