edu.umass.cs.mallet.base.maximize

Class LimitedMemoryBFGS

java.lang.Object

edu.umass.cs.mallet.base.maximize.LimitedMemoryBFGS

All Implemented Interfaces:

Maximizer.ByGradient

public class LimitedMemoryBFGS
extends Object
implements Maximizer.ByGradient

Constructor Summary

Constructors

Constructor and Description
LimitedMemoryBFGS()

Method Summary

Modifier and Type	Method and Description
boolean	maximize(Maximizable.ByGradient maxable)
boolean	maximize(Maximizable.ByGradient maxable, int numIterations)
void	reset() Resets the previous gradients and values that are used to approximate the Hessian.
void	setEvaluator(OptimizerEvaluator eval)
void	setTolerance(double newtol)

Methods inherited from class java.lang.Object

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

Constructor Detail

LimitedMemoryBFGS

public LimitedMemoryBFGS()

Method Detail

setTolerance

public void setTolerance(double newtol)

setEvaluator

public void setEvaluator(OptimizerEvaluator eval)

maximize

public boolean maximize(Maximizable.ByGradient maxable)

Specified by:

maximize in interface Maximizer.ByGradient

maximize

public boolean maximize(Maximizable.ByGradient maxable,
                        int numIterations)

Specified by:

maximize in interface Maximizer.ByGradient

reset

public void reset()

Resets the previous gradients and values that are used to approximate the Hessian. NOTE - If the Maximizable object is modified externally, this method should be called to avoid IllegalStateExceptions.