C# Class Encog.Neural.Networks.Training.PNN.DeriveMinimum

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 Sons Inc (Computers); April 3, 1995 ISBN: 0471105880

ファイルを表示 Open project: encog/encog-silverlight-core Class Usage Examples

Public Methods

Method	Description
Calculate ( int maxIterations, double maxError, double eps, double tol, ICalculationCriteria network, int n, double x, double ystart, double bs, double direc, double g, double h, double deriv2 ) : double	Derive the minimum, using a conjugate gradient method.

Private Methods

Method	Description
FindNewDir ( int n, double gam, double g, double h, double grad ) : void	Find gamma.
Gamma ( int n, double g, double grad ) : double	Find correction for next iteration.

Method Details

Calculate() public method

Derive the minimum, using a conjugate gradient method.

public Calculate ( int maxIterations, double maxError, double eps, double tol, ICalculationCriteria network, int n, double x, double ystart, double bs, double direc, double g, double h, double deriv2 ) : double
maxIterations	int	The max iterations.
maxError	double	Stop at this error rate.
eps	double	The machine's precision.
tol	double	The convergence tolerance.
network	ICalculationCriteria	The network to get the error from.
n	int	The number of variables.
x	double	The independent variable.
ystart	double	The start for y.
bs	double	Work vector, must have n elements.
direc	double	Work vector, must have n elements.
g	double	Work vector, must have n elements.
h	double	Work vector, must have n elements.
deriv2	double	Work vector, must have n elements.
return	double