Class LARSTrainer

java.lang.Object

org.tribuo.regression.slm.SLMTrainer

org.tribuo.regression.slm.LARSTrainer

All Implemented Interfaces:

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

public class LARSTrainer
extends SLMTrainer

A trainer for a linear regression model which uses least angle regression. Each output dimension is trained independently.

See:

 Efron B, Hastie T, Johnstone I, Tibshirani R.
 "Least Angle Regression"
 The Annals of Statistics, 2004.
 

Field Summary

Fields inherited from class org.tribuo.regression.slm.SLMTrainer

maxNumFeatures, normalize, trainInvocationCounter

Fields inherited from interface org.tribuo.Trainer

DEFAULT_SEED, INCREMENT_INVOCATION_COUNT

Constructor Summary

Constructors

Constructor	Description
LARSTrainer()	Constructs a least angle regression trainer that selects all the features.
LARSTrainer(int maxNumFeatures)	Constructs a least angle regression trainer for a linear model.

Method Summary

Modifier and Type	Method	Description
protected DenseVector	newWeights(org.tribuo.regression.slm.SLMTrainer.SLMState state)	Computes the new feature weights.
String	toString()

Methods inherited from class org.tribuo.regression.slm.SLMTrainer

getInvocationCount, getProvenance, setInvocationCount, train, train

Methods inherited from class java.lang.Object

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

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

postConfig

Methods inherited from interface org.tribuo.SparseTrainer

train

Constructor Details

LARSTrainer

public LARSTrainer(int maxNumFeatures)

Constructs a least angle regression trainer for a linear model.

Parameters:

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

LARSTrainer

public LARSTrainer()

Constructs a least angle regression trainer that selects all the features.

Method Details

newWeights

protected DenseVector newWeights(org.tribuo.regression.slm.SLMTrainer.SLMState state)

Description copied from class: SLMTrainer

Computes the new feature weights.

In this version it returns the ordinary least squares solution for the current state.

Overrides:

newWeights in class SLMTrainer

Parameters:

state - The SLM state to operate on.

Returns:

The new feature weights.

toString

public String toString()

Overrides:

toString in class SLMTrainer