Package org.tribuo.regression.xgboost

Class XGBoostRegressionConverter

java.lang.Object

org.tribuo.regression.xgboost.XGBoostRegressionConverter

All Implemented Interfaces:

Serializable, XGBoostOutputConverter<Regressor>, ProtoSerializable<org.tribuo.common.xgboost.protos.XGBoostOutputConverterProto>

public final class XGBoostRegressionConverter
extends Object
implements XGBoostOutputConverter<Regressor>

Converts XGBoost outputs into Regressor Predictions.

Instances of this class are stateless and thread-safe.

See Also:

• Serialized Form

Field Summary

Fields

Modifier and Type	Field	Description
static final int	CURRENT_VERSION	Protobuf serialization version.

Fields inherited from interface org.tribuo.protos.ProtoSerializable

DESERIALIZATION_METHOD_NAME, PROVENANCE_SERIALIZER

Constructor Summary

Constructors

Constructor	Description
XGBoostRegressionConverter()	Construct an XGBoostRegressionConverter.

Method Summary

Modifier and Type	Method	Description
List<Prediction<Regressor>>	convertBatchOutput(ImmutableOutputInfo<Regressor> info, List<float[][]> probabilities, int[] numValidFeatures, Example<Regressor>[] examples)	Converts a list of float arrays from XGBoost Boosters into a Tribuo Prediction.
Prediction<Regressor>	convertOutput(ImmutableOutputInfo<Regressor> info, List<float[]> probabilities, int numValidFeatures, Example<Regressor> example)	Converts a list of float arrays from XGBoost Boosters into a Tribuo Prediction.
static XGBoostRegressionConverter	deserializeFromProto(int version, String className, com.google.protobuf.Any message)	Deserialization factory.
boolean	generatesProbabilities()	Does this converter produce probabilities?
Class<Regressor>	getTypeWitness()	Gets the type witness for the output this converter uses.
org.tribuo.common.xgboost.protos.XGBoostOutputConverterProto	serialize()	Serializes this object to a protobuf.

Methods inherited from class java.lang.Object

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

Field Details

CURRENT_VERSION

public static final int CURRENT_VERSION

Protobuf serialization version.

See Also:

• Constant Field Values

Constructor Details

XGBoostRegressionConverter

public XGBoostRegressionConverter()

Construct an XGBoostRegressionConverter.

Method Details

deserializeFromProto

public static XGBoostRegressionConverter deserializeFromProto(int version,
 String className,
 com.google.protobuf.Any message)

Deserialization factory.

Parameters:

version - The serialized object version.

className - The class name.

message - The serialized data.

Returns:

The deserialized object.

generatesProbabilities

public boolean generatesProbabilities()

Description copied from interface: XGBoostOutputConverter

Does this converter produce probabilities?

Specified by:

generatesProbabilities in interface XGBoostOutputConverter<Regressor>

Returns:

True if it produces probabilities.

convertOutput

public Prediction<Regressor> convertOutput(ImmutableOutputInfo<Regressor> info,
 List<float[]> probabilities,
 int numValidFeatures,
 Example<Regressor> example)

Description copied from interface: XGBoostOutputConverter

Converts a list of float arrays from XGBoost Boosters into a Tribuo Prediction.

Specified by:

convertOutput in interface XGBoostOutputConverter<Regressor>

Parameters:

info - The output info.

probabilities - The XGBoost output.

numValidFeatures - The number of valid features used in the prediction.

example - The example this prediction was generated from.

Returns:

The prediction object.

convertBatchOutput

public List<Prediction<Regressor>> convertBatchOutput(ImmutableOutputInfo<Regressor> info,
 List<float[][]> probabilities,
 int[] numValidFeatures,
 Example<Regressor>[] examples)

Description copied from interface: XGBoostOutputConverter

Converts a list of float arrays from XGBoost Boosters into a Tribuo Prediction.

Specified by:

convertBatchOutput in interface XGBoostOutputConverter<Regressor>

Parameters:

info - The output info.

probabilities - The XGBoost output, list dimension is across models, first array dimension is across examples, second array dimension is across outputs.

numValidFeatures - The number of valid features used in each prediction.

examples - The examples these predictions were generated from.

Returns:

The prediction object.

serialize

public org.tribuo.common.xgboost.protos.XGBoostOutputConverterProto serialize()

Description copied from interface: ProtoSerializable

Serializes this object to a protobuf.

Specified by:

serialize in interface ProtoSerializable<org.tribuo.common.xgboost.protos.XGBoostOutputConverterProto>

Returns:

The protobuf.

getTypeWitness

public Class<Regressor> getTypeWitness()

Description copied from interface: XGBoostOutputConverter

Gets the type witness for the output this converter uses.

The default implementation throws UnsupportedOperationException for compatibility with subclasses which don't support protobuf serialization.

Specified by:

getTypeWitness in interface XGBoostOutputConverter<Regressor>

Returns:

The class of the output.