org.apache.mahout.classifier.sgd
Class AdaptiveLogisticRegression.Wrapper
java.lang.Object
org.apache.mahout.classifier.sgd.AdaptiveLogisticRegression.Wrapper
- All Implemented Interfaces:
- Payload<AdaptiveLogisticRegression.Wrapper>
- Enclosing class:
- AdaptiveLogisticRegression
public static class AdaptiveLogisticRegression.Wrapper
- extends java.lang.Object
- implements Payload<AdaptiveLogisticRegression.Wrapper>
Provides a shim between the EP optimization stuff and the CrossFoldLearner. The most important
interface has to do with the parameters of the optimization. These are taken from the double[]
params in the following order
- regularization constant lambda
- learningRate
.
All other parameters are set in such a way so as to defeat annealing to the extent possible.
This lets the evolutionary algorithm handle the annealing.
Note that per coefficient annealing is still done and no optimization of the per coefficient
offset is done.
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait |
AdaptiveLogisticRegression.Wrapper
public AdaptiveLogisticRegression.Wrapper(int numCategories,
int numFeatures,
PriorFunction prior)
copy
public AdaptiveLogisticRegression.Wrapper copy()
- Specified by:
copy
in interface Payload<AdaptiveLogisticRegression.Wrapper>
update
public void update(double[] params)
- Specified by:
update
in interface Payload<AdaptiveLogisticRegression.Wrapper>
freeze
public void freeze(State<AdaptiveLogisticRegression.Wrapper> s)
setMappings
public void setMappings(State<AdaptiveLogisticRegression.Wrapper> x)
train
public void train(AdaptiveLogisticRegression.TrainingExample example)
getLearner
public CrossFoldLearner getLearner()
toString
public java.lang.String toString()
- Overrides:
toString
in class java.lang.Object
setAucEvaluator
public void setAucEvaluator(OnlineAuc auc)
Copyright © 2008-2010 The Apache Software Foundation. All Rights Reserved.