org.encog.neural.networks.training.strategy
Class SmartMomentum

java.lang.Object
  extended by org.encog.neural.networks.training.strategy.SmartMomentum
All Implemented Interfaces:
Strategy

public class SmartMomentum
extends Object
implements Strategy

Attempt to automatically set a momentum in a training algorithm that supports momentum.

Author:
jheaton

Field Summary
static double MAX_MOMENTUM
          The maximum value that momentum can go to.
static double MIN_IMPROVEMENT
          The minimum improvement to adjust momentum.
static double MOMENTUM_CYCLES
          How many cycles to accept before adjusting momentum.
static double MOMENTUM_INCREASE
          How much to increase momentum by.
static double START_MOMENTUM
          The starting momentum.
 
Constructor Summary
SmartMomentum()
           
 
Method Summary
 void init(MLTrain train)
          Initialize this strategy.
 void postIteration()
          Called just after a training iteration.
 void preIteration()
          Called just before a training iteration.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Field Detail

MIN_IMPROVEMENT

public static final double MIN_IMPROVEMENT
The minimum improvement to adjust momentum.

See Also:
Constant Field Values

MAX_MOMENTUM

public static final double MAX_MOMENTUM
The maximum value that momentum can go to.

See Also:
Constant Field Values

START_MOMENTUM

public static final double START_MOMENTUM
The starting momentum.

See Also:
Constant Field Values

MOMENTUM_INCREASE

public static final double MOMENTUM_INCREASE
How much to increase momentum by.

See Also:
Constant Field Values

MOMENTUM_CYCLES

public static final double MOMENTUM_CYCLES
How many cycles to accept before adjusting momentum.

See Also:
Constant Field Values
Constructor Detail

SmartMomentum

public SmartMomentum()
Method Detail

init

public void init(MLTrain train)
Initialize this strategy.

Specified by:
init in interface Strategy
Parameters:
train - The training algorithm.

postIteration

public void postIteration()
Called just after a training iteration.

Specified by:
postIteration in interface Strategy

preIteration

public void preIteration()
Called just before a training iteration.

Specified by:
preIteration in interface Strategy


Copyright © 2014. All Rights Reserved.