Documentation of 'org.encog.mathutil.matrices.hessian.BasicHessian' Java class
BasicHessian
org.encog.mathutil.matrices.hessian

Class BasicHessian

    • Constructor Detail

      • BasicHessian

        public BasicHessian()
    • Method Detail

      • init

        public void init(BasicNetwork theNetwork,
                         MLDataSet theTraining)
        Init the class.
        Specified by:
        init in interface ComputeHessian
        Parameters:
        theNetwork - The neural network to train.
        theTraining - The training set to train with.
      • getGradients

        public double[] getGradients()
        The gradeints.
        Specified by:
        getGradients in interface ComputeHessian
        Returns:
        The gradients for the Hessian.
      • getHessian

        public double[][] getHessian()
        Specified by:
        getHessian in interface ComputeHessian
        Returns:
        Get the Hessian as a 2d array.
      • clear

        public void clear()
        Clear the Hessian and gradients.
        Specified by:
        clear in interface ComputeHessian
      • getSSE

        public double getSSE()
        Specified by:
        getSSE in interface ComputeHessian
        Returns:
        The sum of squares error over all of the training elements.
      • updateHessian

        public void updateHessian(double[] d)
        Update the Hessian, sum's with what is in the Hessian already. Call clear to clear out old Hessian.
        Parameters:
        d - Vector to update with.

DMelt 3.0 © DataMelt by jWork.ORG

You see the box below because you did not login.