org.apache.spark.ml.classification

Class ProbabilisticClassificationModel<FeaturesType,M extends ProbabilisticClassificationModel<FeaturesType,M>>

java.lang.Object

org.apache.spark.ml.PipelineStage

org.apache.spark.ml.Transformer

org.apache.spark.ml.Model<M>

org.apache.spark.ml.PredictionModel<FeaturesType,M>

org.apache.spark.ml.classification.ClassificationModel<FeaturesType,M>

org.apache.spark.ml.classification.ProbabilisticClassificationModel<FeaturesType,M>

All Implemented Interfaces:

java.io.Serializable, Logging, Params, Identifiable

Direct Known Subclasses:

DecisionTreeClassificationModel, LogisticRegressionModel, NaiveBayesModel, RandomForestClassificationModel

public abstract class ProbabilisticClassificationModel<FeaturesType,M extends ProbabilisticClassificationModel<FeaturesType,M>>
extends ClassificationModel<FeaturesType,M>

:: DeveloperApi ::

Model produced by a ProbabilisticClassifier. Classes are indexed {0, 1, ..., numClasses - 1}.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
ProbabilisticClassificationModel()

Method Summary

Methods

Modifier and Type	Method and Description
static void	normalizeToProbabilitiesInPlace(DenseVector v) Normalize a vector of raw predictions to be a multinomial probability vector, in place.
protected Vector	predictProbability(FeaturesType features) Predict the probability of each class given the features.
protected double	probability2prediction(Vector probability) Given a vector of class conditional probabilities, select the predicted label.
protected double	raw2prediction(Vector rawPrediction) Given a vector of raw predictions, select the predicted label.
protected Vector	raw2probability(Vector rawPrediction) Non-in-place version of raw2probabilityInPlace()
protected abstract Vector	raw2probabilityInPlace(Vector rawPrediction) Estimate the probability of each class given the raw prediction, doing the computation in-place.
M	setProbabilityCol(java.lang.String value)
M	setThresholds(double[] value)
DataFrame	transform(DataFrame dataset) Transforms dataset by reading from featuresCol, and appending new columns as specified by parameters: - predicted labels as predictionCol of type Double - raw predictions (confidences) as rawPredictionCol of type Vector - probability of each class as probabilityCol of type Vector.
StructType	validateAndTransformSchema(StructType schema, boolean fitting, DataType featuresDataType)
StructType	validateAndTransformSchema(StructType schema, boolean fitting, DataType featuresDataType) Validates and transforms the input schema with the provided param map.

Methods inherited from class org.apache.spark.ml.classification.ClassificationModel

numClasses, predict, predictRaw, setRawPredictionCol

Methods inherited from class org.apache.spark.ml.PredictionModel

featuresDataType, setFeaturesCol, setPredictionCol, transformImpl, transformSchema

Methods inherited from class org.apache.spark.ml.Model

copy, hasParent, parent, setParent

Methods inherited from class org.apache.spark.ml.Transformer

transform, transform, transform

Methods inherited from class org.apache.spark.ml.PipelineStage

transformSchema

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.spark.ml.param.Params

clear, copy, copyValues, defaultCopy, defaultParamMap, explainParam, explainParams, extractParamMap, extractParamMap, get, getDefault, getOrDefault, getParam, hasDefault, hasParam, isDefined, isSet, paramMap, params, set, set, set, setDefault, setDefault, shouldOwn, validateParams

Methods inherited from interface org.apache.spark.ml.util.Identifiable

toString, uid

Methods inherited from interface org.apache.spark.Logging

initializeIfNecessary, initializeLogging, isTraceEnabled, log_, log, logDebug, logDebug, logError, logError, logInfo, logInfo, logName, logTrace, logTrace, logWarning, logWarning

Constructor Detail

ProbabilisticClassificationModel

public ProbabilisticClassificationModel()

Method Detail

normalizeToProbabilitiesInPlace

public static void normalizeToProbabilitiesInPlace(DenseVector v)

Normalize a vector of raw predictions to be a multinomial probability vector, in place.

The input raw predictions should be >= 0. The output vector sums to 1, unless the input vector is all-0 (in which case the output is all-0 too).

NOTE: This is NOT applicable to all models, only ones which effectively use class instance counts for raw predictions.

Parameters:

v - (undocumented)

setProbabilityCol

public M setProbabilityCol(java.lang.String value)

setThresholds

public M setThresholds(double[] value)

transform

public DataFrame transform(DataFrame dataset)

Transforms dataset by reading from featuresCol, and appending new columns as specified by parameters: - predicted labels as predictionCol of type Double - raw predictions (confidences) as rawPredictionCol of type Vector - probability of each class as probabilityCol of type Vector.

Overrides:

transform in class ClassificationModel<FeaturesType,M extends ProbabilisticClassificationModel<FeaturesType,M>>

Parameters:

dataset - input dataset

Returns:

transformed dataset

raw2probabilityInPlace

protected abstract Vector raw2probabilityInPlace(Vector rawPrediction)

Estimate the probability of each class given the raw prediction, doing the computation in-place. These predictions are also called class conditional probabilities.

This internal method is used to implement transform() and output probabilityCol.

Parameters:

rawPrediction - (undocumented)

Returns:

Estimated class conditional probabilities (modified input vector)

raw2probability

protected Vector raw2probability(Vector rawPrediction)

Non-in-place version of raw2probabilityInPlace()

raw2prediction

protected double raw2prediction(Vector rawPrediction)

Description copied from class: ClassificationModel

Given a vector of raw predictions, select the predicted label. This may be overridden to support thresholds which favor particular labels.

Overrides:

raw2prediction in class ClassificationModel<FeaturesType,M extends ProbabilisticClassificationModel<FeaturesType,M>>

Parameters:

rawPrediction - (undocumented)

Returns:

predicted label

predictProbability

protected Vector predictProbability(FeaturesType features)

Predict the probability of each class given the features. These predictions are also called class conditional probabilities.

This internal method is used to implement transform() and output probabilityCol.

Parameters:

features - (undocumented)

Returns:

Estimated class conditional probabilities

probability2prediction

protected double probability2prediction(Vector probability)

Given a vector of class conditional probabilities, select the predicted label. This supports thresholds which favor particular labels.

Parameters:

probability - (undocumented)

Returns:

predicted label

validateAndTransformSchema

public StructType validateAndTransformSchema(StructType schema,
                                    boolean fitting,
                                    DataType featuresDataType)

validateAndTransformSchema

public StructType validateAndTransformSchema(StructType schema,
                                    boolean fitting,
                                    DataType featuresDataType)

Validates and transforms the input schema with the provided param map.

Parameters:

schema - input schema

fitting - whether this is in fitting

featuresDataType - SQL DataType for FeaturesType. E.g., VectorUDT for vector features.

Returns:

output schema