org.tribuo.regression.impl

Class SkeletalIndependentRegressionSparseModel

java.lang.Object

org.tribuo.Model<T>

org.tribuo.SparseModel<Regressor>

org.tribuo.regression.impl.SkeletalIndependentRegressionSparseModel

All Implemented Interfaces:

com.oracle.labs.mlrg.olcut.provenance.Provenancable<ModelProvenance>, Serializable

Direct Known Subclasses:

SparseLinearModel

public abstract class SkeletalIndependentRegressionSparseModel
extends SparseModel<Regressor>

A SparseModel which wraps n independent regression models, where n is the size of the MultipleRegressor domain. Each model independently predicts a single regression dimension.

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected String[]	dimensions

Fields inherited from class org.tribuo.Model

ALL_OUTPUTS, BIAS_FEATURE, featureIDMap, generatesProbabilities, name, outputIDInfo, provenance, provenanceOutput

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	SkeletalIndependentRegressionSparseModel(String name, String[] dimensions, ModelProvenance modelProvenance, ImmutableFeatureMap featureMap, ImmutableOutputInfo<Regressor> outputInfo, Map<String,List<String>> activeFeatures) models.size() must equal labelInfo.getDomain().size()

Method Summary

Modifier and Type	Method and Description
protected SparseVector	createFeatures(Example<Regressor> example) Creates the feature vector.
Prediction<Regressor>	predict(Example<Regressor> example) Uses the model to predict the output for a single example.
protected abstract Regressor.DimensionTuple	scoreDimension(int dimensionIdx, SparseVector features) Makes a prediction for a single dimension.

Methods inherited from class org.tribuo.SparseModel

copy, getActiveFeatures

Methods inherited from class org.tribuo.Model

copy, generatesProbabilities, getExcuse, getExcuses, getFeatureIDMap, getName, getOutputIDInfo, getProvenance, getTopFeatures, innerPredict, predict, predict, setName, toString, validate

Methods inherited from class java.lang.Object

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

Field Detail

dimensions

protected final String[] dimensions

Constructor Detail

SkeletalIndependentRegressionSparseModel

protected SkeletalIndependentRegressionSparseModel(String name,
                                                   String[] dimensions,
                                                   ModelProvenance modelProvenance,
                                                   ImmutableFeatureMap featureMap,
                                                   ImmutableOutputInfo<Regressor> outputInfo,
                                                   Map<String,List<String>> activeFeatures)

models.size() must equal labelInfo.getDomain().size()

Parameters:

name - Model name.

dimensions - Dimension names.

modelProvenance - The model provenance.

featureMap - The feature domain used in training.

outputInfo - The output domain used in training.

activeFeatures - The active features in this model.

Method Detail

predict

public Prediction<Regressor> predict(Example<Regressor> example)

Description copied from class: Model

Uses the model to predict the output for a single example.

predict does not mutate the example.

Throws IllegalArgumentException if the example has no features or no feature overlap with the model.

Specified by:

predict in class Model<Regressor>

Parameters:

example - the example to predict.

Returns:

the result of the prediction.

createFeatures

protected SparseVector createFeatures(Example<Regressor> example)

Creates the feature vector. Does not include a bias term.

Designed to be overridden, called by the predict method.

Parameters:

example - The example to convert.

Returns:

The feature vector.

scoreDimension

protected abstract Regressor.DimensionTuple scoreDimension(int dimensionIdx,
                                                           SparseVector features)

Makes a prediction for a single dimension.

Parameters:

dimensionIdx - The dimension index to predict.

features - The features to use.

Returns:

A single dimension prediction.