org.tribuo.regression.slm

Class SLMTrainer

java.lang.Object

org.tribuo.regression.slm.SLMTrainer

All Implemented Interfaces:

com.oracle.labs.mlrg.olcut.config.Configurable, com.oracle.labs.mlrg.olcut.provenance.Provenancable<TrainerProvenance>, SparseTrainer<Regressor>, Trainer<Regressor>, WeightedExamples

Direct Known Subclasses:

LARSLassoTrainer, LARSTrainer

public class SLMTrainer
extends Object
implements SparseTrainer<Regressor>, WeightedExamples

A trainer for a sparse linear regression model. Uses sequential forward selection to construct the model. Optionally can normalize the data first. Each output dimension is trained independently with no shared regularization.

Field Summary

Fields

Modifier and Type	Field and Description
protected int	maxNumFeatures
protected boolean	normalize
protected int	trainInvocationCounter

Fields inherited from interface org.tribuo.Trainer

DEFAULT_SEED

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	SLMTrainer() For OLCUT.
	SLMTrainer(boolean normalize) Constructs a trainer for a sparse linear model using sequential forward selection.
	SLMTrainer(boolean normalize, int maxNumFeatures) Constructs a trainer for a sparse linear model using sequential forward selection.

Method Summary

Modifier and Type	Method and Description
int	getInvocationCount() The number of times this trainer instance has had it's train method invoked.
TrainerProvenance	getProvenance()
protected org.apache.commons.math3.linear.RealVector	newWeights(org.tribuo.regression.slm.SLMTrainer.SLMState state)
String	toString()
SparseLinearModel	train(Dataset<Regressor> examples, Map<String,com.oracle.labs.mlrg.olcut.provenance.Provenance> runProvenance) Trains a sparse linear model.

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.tribuo.SparseTrainer

train

Methods inherited from interface com.oracle.labs.mlrg.olcut.config.Configurable

postConfig

Field Detail

maxNumFeatures

@Config(description="Maximum number of features to use.")
protected int maxNumFeatures

normalize

@Config(description="Normalize the data first.")
protected boolean normalize

trainInvocationCounter

protected int trainInvocationCounter

Constructor Detail

SLMTrainer

public SLMTrainer(boolean normalize,
                  int maxNumFeatures)

Constructs a trainer for a sparse linear model using sequential forward selection.

Parameters:

normalize - Normalizes the data first (i.e., removes the bias term).

maxNumFeatures - The maximum number of features to select. Supply -1 to select all features.

SLMTrainer

public SLMTrainer(boolean normalize)

Constructs a trainer for a sparse linear model using sequential forward selection.

Selects all the features.

Parameters:

normalize - Normalizes the data first (i.e., removes the bias term).

SLMTrainer

protected SLMTrainer()

For OLCUT.

Method Detail

newWeights

protected org.apache.commons.math3.linear.RealVector newWeights(org.tribuo.regression.slm.SLMTrainer.SLMState state)

train

public SparseLinearModel train(Dataset<Regressor> examples,
                               Map<String,com.oracle.labs.mlrg.olcut.provenance.Provenance> runProvenance)

Trains a sparse linear model.

Specified by:

train in interface SparseTrainer<Regressor>

Specified by:

train in interface Trainer<Regressor>

Parameters:

examples - The data set containing the examples.

runProvenance - Training run specific provenance (e.g., fold number).

Returns:

A trained sparse linear model.

getInvocationCount

public int getInvocationCount()

Description copied from interface: Trainer

The number of times this trainer instance has had it's train method invoked.

This is used to determine how many times the trainer's RNG has been accessed to ensure replicability in the random number stream.

Specified by:

getInvocationCount in interface Trainer<Regressor>

Returns:

The number of train invocations.

getProvenance

public TrainerProvenance getProvenance()

Specified by:

getProvenance in interface com.oracle.labs.mlrg.olcut.provenance.Provenancable<TrainerProvenance>

toString

public String toString()

Overrides:

toString in class Object