org.apache.mahout.classifier.sgd
Class RankingGradient

java.lang.Object
  extended by org.apache.mahout.classifier.sgd.RankingGradient
All Implemented Interfaces:
Gradient

public class RankingGradient
extends Object
implements Gradient

Uses the difference between this instance and recent history to get a gradient that optimizes ranking performance. Essentially this is the same as directly optimizing AUC. It isn't expected that this would be used alone, but rather that a MixedGradient would use it and a DefaultGradient together to combine both ranking and log-likelihood goals.


Constructor Summary
RankingGradient(int window)
           
 
Method Summary
 void addToHistory(int actual, Vector instance)
           
 Vector apply(String groupKey, int actual, Vector instance, AbstractVectorClassifier classifier)
           
 Gradient getBaseGradient()
           
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

RankingGradient

public RankingGradient(int window)
Method Detail

apply

public final Vector apply(String groupKey,
                          int actual,
                          Vector instance,
                          AbstractVectorClassifier classifier)
Specified by:
apply in interface Gradient

addToHistory

public void addToHistory(int actual,
                         Vector instance)

getBaseGradient

public Gradient getBaseGradient()


Copyright © 2008–2014 The Apache Software Foundation. All rights reserved.