org.encog.mathutil.matrices.hessian
public class HessianFD extends BasicHessian
Modifier and Type | Field and Description |
---|---|
double |
INITIAL_STEP
The initial step size for dStep.
|
flat, gradients, hessian, hessianMatrix, network, sse, training
Constructor and Description |
---|
HessianFD() |
Modifier and Type | Method and Description |
---|---|
void |
compute()
Compute the Hessian.
|
double[] |
createCoefficients()
Compute finite difference coefficients according to the method provided here:
http://en.wikipedia.org/wiki/Finite_difference_coefficients
|
int |
getPointsPerSide() |
void |
init(BasicNetwork theNetwork,
MLDataSet theTraining)
Init the class.
|
void |
setPointsPerSide(int pointsPerSide)
This specifies the number of points per side, default is 5.
|
clear, getGradients, getHessian, getHessianMatrix, getSSE, updateHessian
public final double INITIAL_STEP
public void init(BasicNetwork theNetwork, MLDataSet theTraining)
init
in interface ComputeHessian
init
in class BasicHessian
theNetwork
- The neural network to train.theTraining
- The training set to train with.public void compute()
public double[] createCoefficients()
public int getPointsPerSide()
public void setPointsPerSide(int pointsPerSide)
pointsPerSide
- The number of points per side.Copyright © 2014. All Rights Reserved.