Examples of org.encog.neural.networks.training.propagation.back.Backpropagation

Package org.encog.neural.networks.training.propagation.back

Examples of org.encog.neural.networks.training.propagation.back.Backpropagation

org.encog.neural.networks.training.propagation.back.Backpropagation
This class implements a backpropagation training algorithm for feed forward neural networks. It is used in the same manner as any other training class that implements the Train interface. Backpropagation is a common neural network training algorithm. It works by analyzing the error of the output of the neural network. Each neuron in the output layer's contribution, according to weight, to this error is determined. These weights are then adjusted to minimize this error. This process continues working its way backwards through the layers of the neural network. This implementation of the backpropagation algorithm uses both momentum and a learning rate. The learning rate specifies the degree to which the weight matrixes will be modified through each iteration. The momentum specifies how much the previous learning iteration affects the current. To use no momentum at all specify zero. One primary problem with backpropagation is that the magnitude of the partial derivative is often detrimental to the training of the neural network. The other propagation methods of Manhatten and Resilient address this issue in different ways. In general, it is suggested that you use the resilient propagation technique for most Encog training tasks over back propagation.

    final double learningRate = holder.getDouble(
        MLTrainFactory.PROPERTY_LEARNING_RATE, false, 0.7);
    final double momentum = holder.getDouble(
        MLTrainFactory.PROPERTY_LEARNING_MOMENTUM, false, 0.3);


    return new Backpropagation((BasicNetwork) method, training,
        learningRate, momentum);
  }

View Full Code Here

  {
    MLDataSet trainingData = new BasicMLDataSet(XOR.XOR_INPUT,XOR.XOR_IDEAL);
    
    BasicNetwork network = NetworkUtil.createXORNetworkUntrained();


    MLTrain bprop = new Backpropagation(network, trainingData, 0.7, 0.9);
    NetworkUtil.testTraining(trainingData,bprop,0.01);
  }

View Full Code Here

    BasicNetwork network1 = NetworkUtil.createXORNetworkUntrained();
    BasicNetwork network2 = NetworkUtil.createXORNetworkUntrained();
    MLDataSet trainingData = new BasicMLDataSet(XOR.XOR_INPUT,XOR.XOR_IDEAL);
    
    // train network 1, no continue
    Backpropagation rprop1 = new Backpropagation(network1,trainingData,0.4,0.4);
    rprop1.iteration();
    rprop1.iteration();
    rprop1.iteration();
    rprop1.iteration();
    
    // train network 2, continue
    Backpropagation rprop2 = new Backpropagation(network2,trainingData,0.4,0.4);
    rprop2.iteration();
    rprop2.iteration();
    TrainingContinuation state = rprop2.pause();
    rprop2 = new Backpropagation(network2,trainingData,0.4,0.4);
    rprop2.resume(state);
    rprop2.iteration();
    rprop2.iteration();
    
    // verify weights are the same
    double[] weights1 = NetworkCODEC.networkToArray(network1);
    double[] weights2 = NetworkCODEC.networkToArray(network2);
    
    Assert.assertEquals(rprop1.getError(), rprop2.getError(), 0.01);
    Assert.assertEquals(weights1.length, weights2.length);
    Assert.assertArrayEquals(weights1, weights2, 0.01);
    
  }

View Full Code Here

  /**
   * {@inheritDoc}
   */
  @Override
  public void createTrainer(final boolean singleThreaded) {
    final Propagation train = new Backpropagation(getNetwork(),
        getTraining(), getLearningRate(), getMomentum());


    if (singleThreaded) {
      train.setThreadCount(1);
    } else {
      train.setThreadCount(0);
    }


    for (final Strategy strategy : getStrategies()) {
      train.addStrategy(strategy);
    }


    setTrain(train);
  }

View Full Code Here


public class BackpropagationFactory implements EnsembleTrainFactory {


  @Override
  public MLTrain getTraining(MLMethod mlMethod, MLDataSet trainingData) {
    return (MLTrain) new Backpropagation((BasicNetwork) mlMethod, trainingData);
  }

View Full Code Here

0 1

TOP

Related Classes of org.encog.neural.networks.training.propagation.back.Backpropagation

org.encog.ensemble.training.BackpropagationFactory

org.encog.examples.neural.benchmark.FahlmanEncoder

org.encog.examples.neural.benchmark.SimpleBenchmark

org.encog.examples.neural.recurrent.elman.ElmanXOR

org.encog.examples.neural.recurrent.jordan.JordanXOR

org.encog.ml.factory.train.BackPropFactory

org.encog.neural.flat.train.prop.TrainFlatNetworkBackPropagation

org.encog.neural.networks.training.concurrent.jobs.BPROPJob

org.encog.neural.networks.training.EvaluateNuguyenWidrow

org.encog.neural.networks.training.propagation.TrainingContinuation

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.