Documentation of 'org.encog.ml.train.strategy.end.StoppingStrategy' Java class
StoppingStrategy
org.encog.ml.train.strategy.end

Class StoppingStrategy

  • All Implemented Interfaces:
    EndTrainingStrategy, Strategy


    public class StoppingStrategy
    extends java.lang.Object
    implements EndTrainingStrategy
    A simple early stopping strategy that halts training when the training set no longer improves.
    • Constructor Detail

      • StoppingStrategy

        public StoppingStrategy(MLDataSet theValidationSet)
      • StoppingStrategy

        public StoppingStrategy(int theAllowedStagnantIterations)
    • Method Detail

      • init

        public void init(MLTrain theTrain)
        Initialize this strategy.
        Specified by:
        init in interface Strategy
        Parameters:
        theTrain - The training algorithm.
      • preIteration

        public void preIteration()
        Called just before a training iteration.
        Specified by:
        preIteration in interface Strategy
      • postIteration

        public void postIteration()
        Called just after a training iteration.
        Specified by:
        postIteration in interface Strategy
      • getStagnantIterations

        public int getStagnantIterations()
      • setStagnantIterations

        public void setStagnantIterations(int stagnantIterations)
      • getAllowedStagnantIterations

        public int getAllowedStagnantIterations()
      • setAllowedStagnantIterations

        public void setAllowedStagnantIterations(int allowedStagnantIterations)
      • isSaveBest

        public boolean isSaveBest()
      • setSaveBest

        public void setSaveBest(boolean saveBest)
      • getMinimumImprovement

        public double getMinimumImprovement()
      • setMinimumImprovement

        public void setMinimumImprovement(double minimumImprovement)

DMelt 3.0 © DataMelt by jWork.ORG

You see the box below because you did not login.