org.encog.neural.networks.training.pnn

Class DeriveMinimum

java.lang.Object

org.encog.neural.networks.training.pnn.DeriveMinimum

public class DeriveMinimum
extends java.lang.Object

This class determines optimal values for multiple sigmas in a PNN kernel. This is done using a CJ (conjugate gradient) method. Some of the algorithms in this class are based on C++ code from: Advanced Algorithms for Neural Networks: A C++ Sourcebook by Timothy Masters John Wiley and Sons Inc (Computers); April 3, 1995 ISBN: 0471105880

Constructor Summary

Constructors

Constructor and Description
DeriveMinimum()

Method Summary

Modifier and Type	Method and Description
double	calculate(int maxIterations, double maxError, double eps, double tol, CalculationCriteria network, int n, double[] x, double ystart, double[] base, double[] direc, double[] g, double[] h, double[] deriv2) Derive the minimum, using a conjugate gradient method.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

DeriveMinimum

public DeriveMinimum()

Method Detail

calculate

public double calculate(int maxIterations,
                        double maxError,
                        double eps,
                        double tol,
                        CalculationCriteria network,
                        int n,
                        double[] x,
                        double ystart,
                        double[] base,
                        double[] direc,
                        double[] g,
                        double[] h,
                        double[] deriv2)

Derive the minimum, using a conjugate gradient method.

Parameters:

maxIterations - The max iterations.

maxError - Stop at this error rate.

eps - The machine's precision.

tol - The convergence tolerance.

network - The network to get the error from.

n - The number of variables.

x - The independent variable.

ystart - The start for y.

base - Work vector, must have n elements.

direc - Work vector, must have n elements.

g - Work vector, must have n elements.

h - Work vector, must have n elements.

deriv2 - Work vector, must have n elements.

Returns:

The best error.