C# Класс Encog.Engine.Network.Flat.ActivationFunctions

The activation functions used by the flat networks.

Показать файл Открыть проект Примеры использования класса

Открытые свойства

Свойство	Тип	Описание
PARAM_NAMES	String[][]

Открытые методы

Метод	Описание
CalculateActivation ( int type, double x, double paras, int xOffset, int xLength, int paramOffset ) : void	Calculate an activation.
CalculateActivationBIPOLAR ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void	Calculate the bipolar activation.
CalculateActivationCOMPETITIVE ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void	Calculate the competitive activation.
CalculateActivationDerivative ( int type, double x, double paras, int paramOffset ) : double	Calculate the derivative of the activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.
CalculateActivationDerivativeBIPOLAR ( double x, double paras, int paramOffset ) : double	Calculate the derivative for bipolar activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.
CalculateActivationDerivativeCOMPETITIVE ( double x, double paras, int paramOffset ) : double	Calculate the derivative for competitive activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.
CalculateActivationDerivativeGAUSSIAN ( double x, double paras, int paramOffset ) : double	Calculate the derivative for gaussian activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.
CalculateActivationDerivativeLINEAR ( double x, double paras, int paramOffset ) : double	Calculate the derivative for linear activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.
CalculateActivationDerivativeLOG ( double x, double paras, int paramOffset ) : double	Calculate the derivative for log activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.
CalculateActivationDerivativeRAMP ( double x, double paras, int paramOffset ) : double	Calculate the derivative for ramp activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.
CalculateActivationDerivativeSIGMOID ( double x, double paras, int paramOffset ) : double	Calculate the derivative for sigmoid activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.
CalculateActivationDerivativeSIN ( double x, double paras, int paramOffset ) : double	Calculate the derivative for sin activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.
CalculateActivationDerivativeSOFTMAX ( double x, double paras, int paramOffset ) : double	Calculate the derivative for softmax activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.
CalculateActivationDerivativeSTEP ( double x, double paras, int paramOffset ) : double	Calculate the derivative for step activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.
CalculateActivationDerivativeTANH ( double x, double paras, int paramOffset ) : double	Calculate the derivative for tanh activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.
CalculateActivationGAUSSIAN ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void	Calculate the gaussian activation.
CalculateActivationLINEAR ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void	Calculate the linear activation.
CalculateActivationLOG ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void	Calculate the log activation.
CalculateActivationRAMP ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void	Calculate the ramp activation.
CalculateActivationSIGMOID ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void	Calculate the sigmoid activation.
CalculateActivationSIN ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void	Calculate the sin activation.
CalculateActivationSOFTMAX ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void	Calculate the softmax activation.
CalculateActivationSTEP ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void	Calculate the step activation.
CalculateActivationTANH ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void	Calculate the tanh activation.
CopyParams ( double source, double target, int index ) : int	Copy parameters. Accounts for variable sized parameters.
GetParams ( int index ) : String[]	Get the parameter names for an activation function.

Приватные методы

Метод	Описание
ActivationFunctions ( ) : Encog.Engine	Private constructor.

Описание методов

CalculateActivation() публичный статический Метод

Calculate an activation.

public static CalculateActivation ( int type, double x, double paras, int xOffset, int xLength, int paramOffset ) : void
type	int	The type of activation.
x	double	The input and output array. Input values are provided and thearray is modified to reflect the output.
paras	double
xOffset	int	The offset into X for where we are calculating.
xLength	int	The length of the array we are calculating.
paramOffset	int	The parameter offset.
Результат	void

CalculateActivationBIPOLAR() публичный статический Метод

Calculate the bipolar activation.

public static CalculateActivationBIPOLAR ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void
x	double	The input and output array. Input values are provided and thearray is modified to reflect the output.
paras	double
xOffset	int	The offset into X for where we are calculating.
xLength	int	The length of the array we are calculating.
paramOffset	int	The parameter offset.
Результат	void

CalculateActivationCOMPETITIVE() публичный статический Метод

Calculate the competitive activation.

public static CalculateActivationCOMPETITIVE ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void
x	double	The input and output array. Input values are provided and thearray is modified to reflect the output.
paras	double
xOffset	int	The offset into X for where we are calculating.
xLength	int	The length of the array we are calculating.
paramOffset	int	The parameter offset.
Результат	void

CalculateActivationDerivative() публичный статический Метод

Calculate the derivative of the activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.

public static CalculateActivationDerivative ( int type, double x, double paras, int paramOffset ) : double
type	int	The type of activation.
x	double	The activation to calculate for.
paras	double
paramOffset	int	The offset the parameters begin at.
Результат	double

CalculateActivationDerivativeBIPOLAR() публичный статический Метод

Calculate the derivative for bipolar activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.

public static CalculateActivationDerivativeBIPOLAR ( double x, double paras, int paramOffset ) : double
x	double	The activation to calculate for.
paras	double
paramOffset	int	The offset the parameters begin at.
Результат	double

CalculateActivationDerivativeCOMPETITIVE() публичный статический Метод

Calculate the derivative for competitive activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.

public static CalculateActivationDerivativeCOMPETITIVE ( double x, double paras, int paramOffset ) : double
x	double	The activation to calculate for.
paras	double
paramOffset	int	The offset the parameters begin at.
Результат	double

CalculateActivationDerivativeGAUSSIAN() публичный статический Метод

Calculate the derivative for gaussian activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.

public static CalculateActivationDerivativeGAUSSIAN ( double x, double paras, int paramOffset ) : double
x	double	The activation to calculate for.
paras	double
paramOffset	int	The offset the parameters begin at.
Результат	double

CalculateActivationDerivativeLINEAR() публичный статический Метод

Calculate the derivative for linear activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.

public static CalculateActivationDerivativeLINEAR ( double x, double paras, int paramOffset ) : double
x	double	The activation to calculate for.
paras	double
paramOffset	int	The offset the parameters begin at.
Результат	double

CalculateActivationDerivativeLOG() публичный статический Метод

Calculate the derivative for log activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.

public static CalculateActivationDerivativeLOG ( double x, double paras, int paramOffset ) : double
x	double	The activation to calculate for.
paras	double
paramOffset	int	The offset the parameters begin at.
Результат	double

CalculateActivationDerivativeRAMP() публичный статический Метод

Calculate the derivative for ramp activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.

public static CalculateActivationDerivativeRAMP ( double x, double paras, int paramOffset ) : double
x	double	The activation to calculate for.
paras	double
paramOffset	int	The offset the parameters begin at.
Результат	double

CalculateActivationDerivativeSIGMOID() публичный статический Метод

Calculate the derivative for sigmoid activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.

public static CalculateActivationDerivativeSIGMOID ( double x, double paras, int paramOffset ) : double
x	double	The activation to calculate for.
paras	double
paramOffset	int	The offset the parameters begin at.
Результат	double

CalculateActivationDerivativeSIN() публичный статический Метод

Calculate the derivative for sin activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.

public static CalculateActivationDerivativeSIN ( double x, double paras, int paramOffset ) : double
x	double	The activation to calculate for.
paras	double
paramOffset	int	The offset the parameters begin at.
Результат	double

CalculateActivationDerivativeSOFTMAX() публичный статический Метод

Calculate the derivative for softmax activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.

public static CalculateActivationDerivativeSOFTMAX ( double x, double paras, int paramOffset ) : double
x	double	The activation to calculate for.
paras	double
paramOffset	int	The offset the parameters begin at.
Результат	double

CalculateActivationDerivativeSTEP() публичный статический Метод

Calculate the derivative for step activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.

public static CalculateActivationDerivativeSTEP ( double x, double paras, int paramOffset ) : double
x	double	The activation to calculate for.
paras	double
paramOffset	int	The offset the parameters begin at.
Результат	double

CalculateActivationDerivativeTANH() публичный статический Метод

Calculate the derivative for tanh activation. It is assumed that the value x, which is passed to this method, was the output from this activation. This prevents this method from having to recalculate the activation, just to recalculate the derivative.

public static CalculateActivationDerivativeTANH ( double x, double paras, int paramOffset ) : double
x	double	The activation to calculate for.
paras	double
paramOffset	int	The offset the parameters begin at.
Результат	double

CalculateActivationGAUSSIAN() публичный статический Метод

Calculate the gaussian activation.

public static CalculateActivationGAUSSIAN ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void
x	double	The input and output array. Input values are provided and thearray is modified to reflect the output.
paras	double
xOffset	int	The offset into X for where we are calculating.
xLength	int	The length of the array we are calculating.
paramOffset	int	The parameter offset.
Результат	void

CalculateActivationLINEAR() публичный статический Метод

Calculate the linear activation.

public static CalculateActivationLINEAR ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void
x	double	The input and output array. Input values are provided and thearray is modified to reflect the output.
paras	double
xOffset	int	The offset into X for where we are calculating.
xLength	int	The length of the array we are calculating.
paramOffset	int	The parameter offset.
Результат	void

CalculateActivationLOG() публичный статический Метод

Calculate the log activation.

public static CalculateActivationLOG ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void
x	double	The input and output array. Input values are provided and thearray is modified to reflect the output.
paras	double
xOffset	int	The offset into X for where we are calculating.
xLength	int	The length of the array we are calculating.
paramOffset	int	The parameter offset.
Результат	void

CalculateActivationRAMP() публичный статический Метод

Calculate the ramp activation.

public static CalculateActivationRAMP ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void
x	double	The input and output array. Input values are provided and thearray is modified to reflect the output.
paras	double
xOffset	int	The offset into X for where we are calculating.
xLength	int	The length of the array we are calculating.
paramOffset	int	The parameter offset.
Результат	void

CalculateActivationSIGMOID() публичный статический Метод

Calculate the sigmoid activation.

public static CalculateActivationSIGMOID ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void
x	double	The input and output array. Input values are provided and thearray is modified to reflect the output.
paras	double
xOffset	int	The offset into X for where we are calculating.
xLength	int	The length of the array we are calculating.
paramOffset	int	The parameter offset.
Результат	void

CalculateActivationSIN() публичный статический Метод

Calculate the sin activation.

public static CalculateActivationSIN ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void
x	double	The input and output array. Input values are provided and thearray is modified to reflect the output.
paras	double
xOffset	int	The offset into X for where we are calculating.
xLength	int	The length of the array we are calculating.
paramOffset	int	The parameter offset.
Результат	void

CalculateActivationSOFTMAX() публичный статический Метод

Calculate the softmax activation.

public static CalculateActivationSOFTMAX ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void
x	double	The input and output array. Input values are provided and thearray is modified to reflect the output.
paras	double
xOffset	int	The offset into X for where we are calculating.
xLength	int	The length of the array we are calculating.
paramOffset	int	The parameter offset.
Результат	void

CalculateActivationSTEP() публичный статический Метод

Calculate the step activation.

public static CalculateActivationSTEP ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void
x	double	The input and output array. Input values are provided and thearray is modified to reflect the output.
paras	double
xOffset	int	The offset into X for where we are calculating.
xLength	int	The length of the array we are calculating.
paramOffset	int	The parameter offset.
Результат	void

CalculateActivationTANH() публичный статический Метод

Calculate the tanh activation.

public static CalculateActivationTANH ( double x, double paras, int xOffset, int xLength, int paramOffset ) : void
x	double	The input and output array. Input values are provided and thearray is modified to reflect the output.
paras	double
xOffset	int	The offset into X for where we are calculating.
xLength	int	The length of the array we are calculating.
paramOffset	int	The parameter offset.
Результат	void

CopyParams() публичный статический Метод

Copy parameters. Accounts for variable sized parameters.

public static CopyParams ( double source, double target, int index ) : int
source	double	The source.
target	double	The target.
index	int	The index.
Результат	int

GetParams() публичный статический Метод

Get the parameter names for an activation function.

public static GetParams ( int index ) : String[]
index	int	The activation function.
Результат	String[]

Описание свойств

PARAM_NAMES публичное статическое свойство

The names of all of the params.

public static String[][] PARAM_NAMES
Результат	String[][]