org.neuroph.nnet.learning
Class SigmoidDeltaRule
java.lang.Object
java.util.Observable
org.neuroph.core.learning.LearningRule
org.neuroph.core.learning.IterativeLearning
org.neuroph.core.learning.SupervisedLearning
org.neuroph.nnet.learning.LMS
org.neuroph.nnet.learning.SigmoidDeltaRule
- All Implemented Interfaces:
- java.io.Serializable, java.lang.Runnable
- Direct Known Subclasses:
- BackPropagation
public class SigmoidDeltaRule
- extends LMS
Delta rule learning algorithm for perceptrons with sigmoid (or any other diferentiable continuous) functions.
TODO: Rename to DeltaRuleContinuous (ContinuousDeltaRule) or something like that, but that will break backward compatibility,
posibly with backpropagation which is the most used
- See Also:
LMS
,
Serialized Form
Method Summary |
protected void |
adjustOutputNeurons(java.util.Vector<java.lang.Double> patternError)
This method implements weights update procedure for the output neurons |
protected void |
updateNetworkWeights(java.util.Vector<java.lang.Double> patternError)
This method implements weight update procedure for the whole network for
this learning rule |
Methods inherited from class org.neuroph.core.learning.SupervisedLearning |
doLearningEpoch, errorChangeStalled, getMaxError, getMinErrorChange, getMinErrorChangeIterationsCount, getMinErrorChangeIterationsLimit, getPatternError, getPreviousEpochError, getTotalNetworkError, hasReachedStopCondition, learn, learn, learnPattern, reset, setMaxError, setMinErrorChange, setMinErrorChangeIterationsLimit |
Methods inherited from class java.util.Observable |
addObserver, clearChanged, countObservers, deleteObserver, deleteObservers, hasChanged, notifyObservers, notifyObservers, setChanged |
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
SigmoidDeltaRule
public SigmoidDeltaRule()
- Creates new SigmoidDeltaRule
SigmoidDeltaRule
public SigmoidDeltaRule(NeuralNetwork neuralNetwork)
- Creates new SigmoidDeltaRule for the specified neural network
- Parameters:
neuralNetwork
- neural network to train
updateNetworkWeights
protected void updateNetworkWeights(java.util.Vector<java.lang.Double> patternError)
- This method implements weight update procedure for the whole network for
this learning rule
- Overrides:
updateNetworkWeights
in class LMS
- Parameters:
patternError
- single pattern error vector
adjustOutputNeurons
protected void adjustOutputNeurons(java.util.Vector<java.lang.Double> patternError)
- This method implements weights update procedure for the output neurons
- Parameters:
patternError
- single pattern error vector