MultiLayerPerceptron

java.lang.Object

operations.TrainingNetwork

network.MultiLayerPerceptron

public class MultiLayerPerceptron extends TrainingNetwork

Constructor Summary

Constructors

Constructor

Description

MultiLayerPerceptron(NeuralNetwork neuralNetwork)

MultiLayerPerceptron(TypeOfLearning typeOfLearning, TrainingStrategy trainingStrategies, NeuralNetwork neuralNetwork, InputSamples inputSamples, Report report)
Method Summary

Modifier and Type

Method

Description

void

backPropagation()

O método backPropagation se manifesta a cada período estabelecido para, com as informações dos erros, reconfigurar os pessos das Connections.

void

feedForward()

O método feedForward é responsável por acompanhar o resultado de cada NeuralNetwork.propagate(), comparar os resultados alcançados, calcular os erros, adicioná-los ao Array de erros e o Array de tempo gasto, a serem passados para o Report.

void

startTraining()

O método startTraining() é responsável por conduzir as chamadas de TrainingNetwork.feedForward() e TrainingNetwork.backPropagation() de tal forma que a rede possa ser treinada corretamente.

Methods inherited from class operations.TrainingNetwork
getReport, getTrainingStrategies, getTypeOfLearning, setReport, setTrainingStrategies, setTypeOfLearning

Methods inherited from class java.lang.Object
equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details
- MultiLayerPerceptron
  
  public MultiLayerPerceptron(NeuralNetwork neuralNetwork)
- MultiLayerPerceptron
  
  public MultiLayerPerceptron(TypeOfLearning typeOfLearning, TrainingStrategy trainingStrategies, NeuralNetwork neuralNetwork, InputSamples inputSamples, Report report)
Method Details
- feedForward
  
  public void feedForward()
  
  Description copied from class: TrainingNetwork
  
  O método feedForward é responsável por acompanhar o resultado de cada NeuralNetwork.propagate(), comparar os resultados alcançados, calcular os erros, adicioná-los ao Array de erros e o Array de tempo gasto, a serem passados para o Report.
  
  Specified by:
  
  feedForward in class TrainingNetwork
- backPropagation
  
  public void backPropagation()
  
  Description copied from class: TrainingNetwork
  
  O método backPropagation se manifesta a cada período estabelecido para, com as informações dos erros, reconfigurar os pessos das Connections.
  
  Specified by:
  
  backPropagation in class TrainingNetwork
- startTraining
  
  public void startTraining()
  
  Description copied from class: TrainingNetwork
  
  O método startTraining() é responsável por conduzir as chamadas de TrainingNetwork.feedForward() e TrainingNetwork.backPropagation() de tal forma que a rede possa ser treinada corretamente.
  
  Specified by:
  
  startTraining in class TrainingNetwork