- All Implemented Interfaces:
public class NeuralSimulatedAnnealingextends BasicTrainingThis class implements a simulated annealing training algorithm for neural networks. It is based on the generic SimulatedAnnealing class. It is used in the same manner as any other training class that implements the Train interface. There are essentially two ways you can make use of this class. Either way, you will need a score object. The score object tells the simulated annealing algorithm how well suited a neural network is. If you would like to use simulated annealing with a training set you should make use TrainingSetScore class. This score object uses a training set to score your neural network. If you would like to be more abstract, and not use a training set, you can create your own implementation of the CalculateScore method. This class can then score the networks any way that you like.
Fields Modifier and Type Field and Description
CUTThe cutoff for random data.
Constructors Constructor and Description
NeuralSimulatedAnnealing(BasicNetwork network, CalculateScore calculateScore, double startTemp, double stopTemp, int cycles)Construct a simulated annleaing trainer for a feedforward neural network.
Methods Modifier and Type Method and Description
getArray()Get the network as an array of doubles.
getMethod()Get the current best machine learning method from the training.
iteration()Perform one iteration of simulated annealing.
pause()Pause the training to continue later.
putArray(double array)Convert an array of doubles to the current best network.
randomize()Randomize the weights and bias values.
resume(TrainingContinuation state)Resume training.
Methods inherited from class org.encog.ml.train.BasicTraining
addStrategy, finishTraining, getError, getImplementationType, getIteration, getStrategies, getTraining, isTrainingDone, iteration, postIteration, preIteration, setError, setIteration, setTraining
public static final double CUTThe cutoff for random data.
- See Also:
- Constant Field Values
public NeuralSimulatedAnnealing(BasicNetwork network, CalculateScore calculateScore, double startTemp, double stopTemp, int cycles)Construct a simulated annleaing trainer for a feedforward neural network.
network- The neural network to be trained.
calculateScore- Used to calculate the score for a neural network.
startTemp- The starting temperature.
stopTemp- The ending temperature.
cycles- The number of cycles in a training iteration.
public final boolean canContinue()
- True if the training can be paused, and later continued.
public final double getArray()Get the network as an array of doubles.
- The network as an array of doubles.
public final double getArrayCopy()
- A copy of the annealing array.
public final CalculateScore getCalculateScore()
- The object used to calculate the score.
public final BasicNetwork getMethod()Get the current best machine learning method from the training.
- The best machine learningm method.
public final void iteration()Perform one iteration of simulated annealing.
public TrainingContinuation pause()Description copied from interface:
MLTrainPause the training to continue later.
- A training continuation object.
public final void putArray(double array)Convert an array of doubles to the current best network.
array- An array.
public final void randomize()Randomize the weights and bias values. This function does most of the work of the class. Each call to this class will randomize the data according to the current temperature. The higher the temperature the more randomness.
public void resume(TrainingContinuation state)Resume training.
state- The training continuation object to use to continue.
SCaVis 2.0 © jWork.ORG