Package cc.mallet.optimize

Class StochasticMetaAscent

java.lang.Object

cc.mallet.optimize.StochasticMetaAscent

All Implemented Interfaces:

Optimizer.ByBatches

public class StochasticMetaAscent
extends java.lang.Object
implements Optimizer.ByBatches

Author:

Greg Druck, Kedar Bellare

Constructor Summary

Constructors

Constructor	Description
StochasticMetaAscent(Optimizable.ByBatchGradient maxable)

Method Summary

Modifier and Type	Method	Description
boolean	optimize(int numBatches, int[] batchAssignments)
boolean	optimize(int numIterations, int numBatches, int[] batchAssignments)
void	setInitialStep(double step)
void	setMu(double m)
void	setUseHessian(boolean flag)

Methods inherited from class java.lang.Object

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

Constructor Detail

StochasticMetaAscent

public StochasticMetaAscent(Optimizable.ByBatchGradient maxable)

Method Detail

setInitialStep

public void setInitialStep(double step)

setMu

public void setMu(double m)

setUseHessian

public void setUseHessian(boolean flag)

optimize

public boolean optimize(int numBatches,
                        int[] batchAssignments)

Specified by:

optimize in interface Optimizer.ByBatches

optimize

public boolean optimize(int numIterations,
                        int numBatches,
                        int[] batchAssignments)

Specified by:

optimize in interface Optimizer.ByBatches