org.apache.mahout.classifier.sgd

Class DefaultGradient

java.lang.Object

org.apache.mahout.classifier.sgd.DefaultGradient

All Implemented Interfaces:

Gradient

public class DefaultGradient
extends Object
implements Gradient

Implements the basic logistic training law.

Constructor Summary

Constructors

Constructor and Description
DefaultGradient()

Method Summary

Modifier and Type	Method and Description
Vector	apply(String groupKey, int actual, Vector instance, AbstractVectorClassifier classifier) Provides a default gradient computation useful for logistic regression.

Methods inherited from class java.lang.Object

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

Constructor Detail

DefaultGradient

public DefaultGradient()

Method Detail

apply

public final Vector apply(String groupKey,
                          int actual,
                          Vector instance,
                          AbstractVectorClassifier classifier)

Provides a default gradient computation useful for logistic regression.

Specified by:

apply in interface Gradient

Parameters:

groupKey - A grouping key to allow per-something AUC loss to be used for training.

actual - The target variable value.

instance - The current feature vector to use for gradient computation

classifier - The classifier that can compute scores

Returns:

The gradient to be applied to beta