Contains Parameters that define neural network architecture limits and trainig parameter limits. More...

Collaboration diagram for IG.Num.NeuralTrainingLimits:

Public Member Functions
void	PrepareNumHiddenNeuronsValuesArray (int numLayers, params int[] numHiddenNeuronsValues)
	Prepares values of numbers of neurons in individual layers according to parameters. More...

void	PrepareNumHiddenNeuronsValuesArray (int numLayers, int minNumNeurons, int maxNumNeurons, int numValues, double intervalGrowthFactor)
	Prepares values of numbers of neurons in individual layers according to parameters in such a way that intervals between these values grow exponentially. More...

delegate void	DoForParameters (List< NeuralTrainingParameters > trainingParameters, List< int > dimensions, int which)

void	DoForParameters_CreateTable (List< NeuralTrainingParameters > trainingParameters, List< int > dimensions, int which)

void	CreateTrainingTable (List< NeuralTrainingParameters > trainingParameters, List< int > tableDimensions)
	Creates a table of training parameters. More...

void	CreateTrainingTable (List< NeuralTrainingParameters > trainingParameters)
	Creates a table of training parameters. More...

void	IterateThroughMultidimensionalTable (List< NeuralTrainingParameters > tableResults, List< int > tableDimensions, bool createTable, bool createDimensions, DoForParameters doOnElement)
	Iterates through all elements of the table of training results defined by the current object, and does whatever is specified by the parameters. More...

Static Public Member Functions
static void	SaveJson (NeuralTrainingLimits trainingLimits, string filePath)
	Saves (serializes) the specified array of training limits objects to the specified JSON file. File is owerwritten if it exists. More...

static void	SaveJson (NeuralTrainingLimits trainingLimits, string filePath, bool append)
	Saves (serializes) the specified array of training limits objects to the specified JSON file. More...

static void	LoadJson (string filePath, ref NeuralTrainingLimits trainingLimits)
	Restores (deserializes) an array of training parameters objects from the specified file in JSON format. More...

Static Public Attributes
static double	DefaultLearningRateMin = 0.1
	Default value for minimum learning rate in neural networks. More...

static double	DefaultLearningRateMax = 0.6
	Default value for maximum learning rate in neural networks. More...

static int	DefaultLearningRateNum = 5
	Default value for number of learning rates in neural networks. More...

static double	DefaultMomentumMin = 0.3
	Default value for minimum momentum in neural networks. More...

static double	DefaultMomentumMax = 0.8
	Default value for maximum momentum in neural networks. More...

static int	DefaultMomentumNum = 5
	Default value for number of momentums in neural networks. More...

static double	DefaultAlphaMin = 1.0
	Default value for minimum alpha value in neural networks. More...

static double	DefaultAlphaMax = 2.0
	Default value for maximum alpha value in neural networks. More...

static int	DefaultAlphaNum = 5
	Default value for number of alpha values in neural networks. More...

static double	DefaultInputSafetyFactorMin = 1.4
	Default value for minimum input bound safety factor value in neural networks. More...

static double	DefaultInputSafetyFactorMax = 1.4
	Default value for maximum input bound safety factor value in neural networks. More...

static int	DefaultInputSafetyFactorNum = 1
	Default value for number of input bound safety factors values in neural networks. More...

static double	DefaultOutputSafetyFactorMin = 1.4
	Default value for minimum output bound safety factor value in neural networks. More...

static double	DefaultOutputSafetyFactorMax = 1.4
	Default value for maximum output bound safety factor value in neural networks. More...

static int	DefaultOutputSafetyFactorNum = 1
	Default value for number of output bound safety factors values in neural networks. More...

static bool	DefaultEnableArchitectureTest = false
	Default flag for enabling different layers in neural networks. More...

static int	DefaultNumHiddenLayersNum = 1
	Default value for number of hidden layers in neural networks. More...

static int	DefaultNumHiddenNeuronsFirstMin = 5
	Default value for number of hidden neurons in first hidden layer in neural networks. More...

static int	DefaultNumHiddenNeuronsFirstMax = 25
	Default value for number of hidden neurons in first hidden layer in neural networks. More...

static int	DefaultNumHiddenNeuronsFirstNum = 1
	Default value for number of hidden neurons in first hidden layer in neural networks. More...

static int	DefaultNumHiddenNeuronsSecondMin = 2
	Default value for number of hidden neurons in second hidden layer in neural networks. More...

static int	DefaultNumHiddenNeuronsSecondMax = 10
	Default value for number of hidden neurons in second hidden layer in neural networks. More...

static int	DefaultNumHiddenNeuronsSecondNum = 1
	Default value for number of hidden neurons in second hidden layer in neural networks. More...

static int	DefaultNumHiddenNeuronsThirdMin = 2
	Default value for number of hidden neurons in third hidden layer in neural networks. More...

static int	DefaultNumHiddenNeuronsThirdMax = 10
	Default value for number of hidden neurons in third hidden layer in neural networks. More...

static int	DefaultNumHiddenNeuronsThirdNum = 1
	Default value for number of hidden neurons in third hidden layer in neural networks. More...

static int	DefaultMaxEpochs = 1000
	Default value for maximum number of epochs in neural networks. More...

static int	DefaultEpochBundle = 100
	Default value for number of epochs in bundle in neural networks. More...

static bool	DefaultEnableRangeTolerance = false
	Flag for enabling toelrance that represent a percentage of the output range. More...

static int	DefaultInputLenght = 5
	Default value for number of input neurons. More...

static int	DefaultOutputLenght = 3
	Default value for number of output neurons. More...

Protected Member Functions
int[]	GetArrayCopy (int[] original)
	Creates and returns a copy of the specified array of integers. More...

int[]	GetArrayCopyInt (double[] original)
	Creates and returns an array of integers that is a copy of the specified array of double values (double to integer conversion made by rounding). More...

Protected Attributes
double	_learningRateMin = NeuralTrainingParameters.DefaultLearningRate

double	_learningRateMax = NeuralTrainingParameters.DefaultLearningRate

int	_learningRateNum = 1

double	_momentumMin = DefaultMomentumMin

double	_momentumMax = DefaultMomentumMax

int	_momentumNum = DefaultMomentumNum

double	_alphaMin = DefaultAlphaMin

double	_alphaMax = DefaultAlphaMax

int	_alphaNum = DefaultAlphaNum

double	_inputSafetyFactorMin = DefaultInputSafetyFactorMin

double	_inputSafetyFactorMax = DefaultInputSafetyFactorMax

int	_inputSafetyFactorNum = DefaultInputSafetyFactorNum

double	_outputSafetyFactorMin = DefaultOutputSafetyFactorMin

double	_outputSafetyFactorMax = DefaultOutputSafetyFactorMax

int	_outputSafetyFactorNum = DefaultOutputSafetyFactorNum

int	_maxEpochs = DefaultMaxEpochs

int	_epochBundle = DefaultEpochBundle

int	_inputLenght = DefaultInputLenght

int	_outputLenght = DefaultOutputLenght

bool	_enableRangeTolerance = DefaultEnableRangeTolerance

IVector	_toleranceRms

IVector	_toleranceMax

bool	_enableArchitectureTest = DefaultEnableArchitectureTest

int	_numHiddenLayersNum = DefaultNumHiddenLayersNum

int	_numHiddenNeuronsFirstMin = DefaultNumHiddenNeuronsFirstMin

int	_numHiddenNeuronsFirstMax = DefaultNumHiddenNeuronsFirstMax

int	_numHiddenNeuronsFirstNum = DefaultNumHiddenNeuronsFirstNum

int[]	_numHiddenNeuronsFirstValues

bool	randomTableEnabled = false

int	_numHiddenNeuronsSecondMin = DefaultNumHiddenNeuronsSecondMin

int	_numHiddenNeuronsSecondMax = DefaultNumHiddenNeuronsSecondMax

int	_numHiddenNeuronsSecondNum = DefaultNumHiddenNeuronsSecondNum

int[]	_numHiddenNeuronsSecondValues

int	_numHiddenNeuronsThirdMin = DefaultNumHiddenNeuronsThirdMin

int	_numHiddenNeuronsThirdMax = DefaultNumHiddenNeuronsThirdMax

int	_numHiddenNeuronsThirdNum = DefaultNumHiddenNeuronsThirdNum

int[]	_numHiddenNeuronsThirdValues

Properties
double	LearningRateMin `[get, set]`
	Minimum limit for learning rate. More...

double	LearningRateMax `[get, set]`
	Maximum limit for learning rate. More...

int	LearningRateNum `[get, set]`
	Number of learning rates. More...

double	MomentumMin `[get, set]`
	Minimum limit for momentum. More...

double	MomentumMax `[get, set]`
	Maximum limit for momentum. More...

int	MomentumNum `[get, set]`
	Number of momentums. More...

double	AlphaMin `[get, set]`
	Minimum limit for alpha value. More...

double	AlphaMax `[get, set]`
	Maximum limit for alpha value. More...

int	AlphaNum `[get, set]`
	Number of alpha value. More...

double	InputSafetyFactorMin `[get, set]`
	Minimum limit for input safety factor value. More...

double	InputSafetyFactorMax `[get, set]`
	Maximum limit for input safety factor value. More...

int	InputSafetyFactorNum `[get, set]`
	Number of input safety factor values. More...

double	OutputSafetyFactorMin `[get, set]`
	Minimum limit for output safety factor value. More...

double	OutputSafetyFactorMax `[get, set]`
	Maximum limit for output safety factor value. More...

int	OutputSafetyFactorNum `[get, set]`
	Number of output safety factor values. More...

int	MaxEpochs `[get, set]`
	Maximum number of epochs performed in training. More...

int	EpochBundle `[get, set]`
	Number of epochs in boundle. More...

int	InputLenght `[get, set]`
	Number of input neurons. More...

int	OutputLength `[get, set]`
	Number of output neurons. More...

bool	EnableRangeTolerance `[get, set]`
	Flag for enabling toelrance that represent a percentage of the output range. More...

IVector	ToleranceRms `[get, set]`
	Tolerance for RMS. More...

IVector	ToleranceMax `[get, set]`
	Maximum tolerance for max. abs. difference. More...

IVector	ToleranceRmsRelativeToRange `[get, set]`
	Relative tolerances on RMS errors of outputs over training points, relative to the correspoinding ranges of output data. More...

double	ToleranceRmsRelativeToRangeScalar `[get, set]`
	Scalar through which all components of the Relative tolerances on RMS errors of outputs can be set to the same value. More...

IVector	ToleranceMaxRelativeToRange `[get, set]`
	Relative tolerances on max. abs. errors of outputs over training points, relative to the correspoinding ranges of output data. More...

double	ToleranceMaxRelativeToRangeScalar `[get, set]`
	Scalar through which all components of the Relative tolerances on max. abs. errors of outputs can be set to the same value. More...

bool	EnableArchitectureTest `[get, set]`
	Flag for enabling test in architecture of ANN. More...

int	NumHiddenLayersNum `[get, set]`
	Number of hidden layers in neural network. More...

int	NumHiddenNeuronsFirstMin `[get, set]`
	Minimum number of hidden neurons in first hidden layer. More...

int	NumHiddenNeuronsFirstMax `[get, set]`
	Maximum number of hidden neurons in first hidden layer. More...

int	NumHiddenNeuronsFirstNum `[get, set]`
	Number of hidden neurons in first hidden layer. More...

int[]	NumHiddenNeuronsFirstValues `[get, set]`
	Values of number of hidden neurons in the first layers that will appear in the table. More...

int	NumHiddenNeuronsSecondMin `[get, set]`
	Minimum number of hidden neurons in second hidden layer. More...

int	NumHiddenNeuronsSecondMax `[get, set]`
	Maximum number of hidden neurons in second hidden layer. More...

int	NumHiddenNeuronsSecondNum `[get, set]`
	Number of hidden neurons in second hidden layer. More...

int[]	NumHiddenNeuronsSecondValues `[get, set]`
	Values of number of hidden neurons in the second layers that will appear in the table. More...

int	NumHiddenNeuronsThirdMin `[get, set]`
	Minimum number of hidden neurons in third hidden layer. More...

int	NumHiddenNeuronsThirdMax `[get, set]`
	Maximum number of hidden neurons in third hidden layer. More...

int	NumHiddenNeuronsThirdNum `[get, set]`
	Number of hidden neurons in third hidden layer. More...

int[]	NumHiddenNeuronsThirdValues `[get, set]`
	Values of number of hidden neurons in the third layer that will appear in the table. More...

Private Member Functions
void	PrepareNeuronsTable (int minNeurons, int maxNeurons, int numNeurons, double growthFactorref, ref int[] neurpnsTable)
	Prepares table of neurons in geometric sequence. More...

void	PrepareNeuronTable (ref int[] neuronsFirstLayer, ref int[] neuronsSecondLayer, ref int[] neuronsThirdLayer)

Private Attributes
IVector	_tolRmsRelative
	Auxiliary properties for defining tolerances in a relative way: More...

double	_tolRmsRelativeScalar = NeuralTrainingParameters.DefaultToleranceRmsRelativeToRangeScalar

IVector	_tolMaxRelative

double	_tolMaxRelativeScalar = NeuralTrainingParameters.DefaultToleranceMaxRelativeToRangeScalar

Detailed Description

Contains Parameters that define neural network architecture limits and trainig parameter limits.

Not thread safe!

This class is used for storing training parameter limits of neural networks and restoring them at a later time, in order to repeat training under similar condition or simply to analyse performance of neural networks.

$A Igor Jul10 Aug12 Nov12; Tako78 Aug12;

Member Function Documentation

int [] IG.Num.NeuralTrainingLimits.GetArrayCopy ( int[] original )

inlineprotected

Creates and returns a copy of the specified array of integers.

Parameters

original Array whose copy is returned.

int [] IG.Num.NeuralTrainingLimits.GetArrayCopyInt ( double[] original )

inlineprotected

Creates and returns an array of integers that is a copy of the specified array of double values (double to integer conversion made by rounding).

Parameters

original Array whose copy is returned.

void IG.Num.NeuralTrainingLimits.PrepareNumHiddenNeuronsValuesArray	(	int	numLayers,
		params int[]	numHiddenNeuronsValues
	)

inline

Prepares values of numbers of neurons in individual layers according to parameters.

For each layer, numbers of neurons in that layers to be used in the table are the same.

Parameters

numLayers	Number of hidden layers (this is fixed for the tables where this method is used; the containing class also does not support tables where number of layers would vary).
numHiddenNeuronsValues	Values for the number of hidden neurons in different layers (common for all layers).

void IG.Num.NeuralTrainingLimits.PrepareNumHiddenNeuronsValuesArray	(	int	numLayers,
		int	minNumNeurons,
		int	maxNumNeurons,
		int	numValues,
		double	intervalGrowthFactor
	)

inline

Prepares values of numbers of neurons in individual layers according to parameters in such a way that intervals between these values grow exponentially.

Parameters

numLayers	Number of hidden layers (this is fixed for the tables where this method is used; the containing class also does not support tables where number of layers would vary).
minNumNeurons	Minimal value for the number fo neurons in hidden layers.
maxNumNeurons	Maxmial value for the number of neurons in hidden layers.
numValues	Number of values for the number of neurons in hidden layers.
intervalGrowthFactor	Factor by which length of each next iterval befoore successive values for the number of neurons in hidden layers is extended.

References IG.Num.GridGenerator1dBase.GetNodeTable().

delegate void IG.Num.NeuralTrainingLimits.DoForParameters	(	List< NeuralTrainingParameters >	trainingParameters,
		List< int >	dimensions,
		int	which
	)

void IG.Num.NeuralTrainingLimits.DoForParameters_CreateTable	(	List< NeuralTrainingParameters >	trainingParameters,
		List< int >	dimensions,
		int	which
	)

inline

void IG.Num.NeuralTrainingLimits.CreateTrainingTable	(	List< NeuralTrainingParameters >	trainingParameters,
		List< int >	tableDimensions
	)

inline

Creates a table of training parameters.

Parameters

trainingParameters	List where table of parameters is stored. Must be different than null. Eventual elements already contained will be deleted.
tableDimensions	List of dimensions of the generated table. If specified then table dimensions will be put on the list (the list is cleared at the beinning).

void IG.Num.NeuralTrainingLimits.CreateTrainingTable ( List< NeuralTrainingParameters > trainingParameters )

inline

Creates a table of training parameters.

Parameters

trainingParameters List where table of parameters is stored. Must be different than null. Eventual elements already contained will be deleted.

void IG.Num.NeuralTrainingLimits.PrepareNeuronsTable	(	int	minNeurons,
		int	maxNeurons,
		int	numNeurons,
		double	growthFactorref,
		ref int[]	neurpnsTable
	)

inlineprivate

Prepares table of neurons in geometric sequence.

Parameters

minNeurons	Minimum number of neurons.
maxNeurons	Maximum number of neurons.
numNeurons	Number of neurons in sequence.
growthFactorref	Growth factor.
neurpnsTable	Table of neurons.

$A Tako78 Nov12;

References IG.Num.GridGenerator1dBase.CoordinateFirst, IG.Num.GridGenerator1dBase.CoordinateLast, IG.Num.GridGenerator1dBase.GetNodeTable(), IG.Num.GridGenerator1d.GrowthFactor, and IG.Num.GridGenerator1dBase.NumNodes.

void IG.Num.NeuralTrainingLimits.PrepareNeuronTable	(	ref int[]	neuronsFirstLayer,
		ref int[]	neuronsSecondLayer,
		ref int[]	neuronsThirdLayer
	)

inlineprivate

void IG.Num.NeuralTrainingLimits.IterateThroughMultidimensionalTable	(	List< NeuralTrainingParameters >	tableResults,
		List< int >	tableDimensions,
		bool	createTable,
		bool	createDimensions,
		DoForParameters	doOnElement
	)

inline

Iterates through all elements of the table of training results defined by the current object, and does whatever is specified by the parameters.

Parameters

tableResults	Table where training parameters and results are stored. Must be specified when this table is created (i.e. parameter createTable is true) or when something must be done on each element (i.e. function parameter is doOnElement defined).
tableDimensions	Table where dimensions are stored. Must be specified when the table of dimensions will be created (i.e. the parameter createDimensions is true).
createTable	Specifies whether the table of training parameters and results will be created by the function call. If true then the parameter tableResults must be specified.
createDimensions	Specifies whether the table of dimensions will be created when iterating over elements. If true then the parameter tableDimensions must be specified.
doOnElement	Delegate that is executed on each element of the table. Index of the current element, as well as the table of the resulet (parameter tableResults ) and the table of dimensions (parameter tableDimensions ) are passed to the delegate when executed.

Referenced by IG.Num.NeuralTrainingTable.UpdateTrainingTableDimensionns().

static void IG.Num.NeuralTrainingLimits.SaveJson	(	NeuralTrainingLimits	trainingLimits,
		string	filePath
	)

inlinestatic

Saves (serializes) the specified array of training limits objects to the specified JSON file. File is owerwritten if it exists.

Parameters

trainingLimits	Contains neural network architectural limits and limits for training parameters.
filePath	Path to the file into which object is is saved.

static void IG.Num.NeuralTrainingLimits.SaveJson	(	NeuralTrainingLimits	trainingLimits,
		string	filePath,
		bool	append
	)

inlinestatic

Saves (serializes) the specified array of training limits objects to the specified JSON file.

Parameters

trainingLimits	Contains neural network architectural limits and limits for training parameters.
filePath	Path to the file into which object is is saved.
append	Specifies whether serialized data is appended at the end of the file in the case that the file already exists.

static void IG.Num.NeuralTrainingLimits.LoadJson	(	string	filePath,
		ref NeuralTrainingLimits	trainingLimits
	)

inlinestatic

Restores (deserializes) an array of training parameters objects from the specified file in JSON format.

Parameters

filePath	File from which array of objects is restored.
trainingLimits	Contains neural network architectural limits and limits for training parameters.

Member Data Documentation

double IG.Num.NeuralTrainingLimits.DefaultLearningRateMin = 0.1

static

Default value for minimum learning rate in neural networks.

double IG.Num.NeuralTrainingLimits.DefaultLearningRateMax = 0.6

static

Default value for maximum learning rate in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultLearningRateNum = 5

static

Default value for number of learning rates in neural networks.

double IG.Num.NeuralTrainingLimits.DefaultMomentumMin = 0.3

static

Default value for minimum momentum in neural networks.

double IG.Num.NeuralTrainingLimits.DefaultMomentumMax = 0.8

static

Default value for maximum momentum in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultMomentumNum = 5

static

Default value for number of momentums in neural networks.

double IG.Num.NeuralTrainingLimits.DefaultAlphaMin = 1.0

static

Default value for minimum alpha value in neural networks.

double IG.Num.NeuralTrainingLimits.DefaultAlphaMax = 2.0

static

Default value for maximum alpha value in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultAlphaNum = 5

static

Default value for number of alpha values in neural networks.

double IG.Num.NeuralTrainingLimits.DefaultInputSafetyFactorMin = 1.4

static

Default value for minimum input bound safety factor value in neural networks.

double IG.Num.NeuralTrainingLimits.DefaultInputSafetyFactorMax = 1.4

static

Default value for maximum input bound safety factor value in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultInputSafetyFactorNum = 1

static

Default value for number of input bound safety factors values in neural networks.

double IG.Num.NeuralTrainingLimits.DefaultOutputSafetyFactorMin = 1.4

static

Default value for minimum output bound safety factor value in neural networks.

double IG.Num.NeuralTrainingLimits.DefaultOutputSafetyFactorMax = 1.4

static

Default value for maximum output bound safety factor value in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultOutputSafetyFactorNum = 1

static

Default value for number of output bound safety factors values in neural networks.

bool IG.Num.NeuralTrainingLimits.DefaultEnableArchitectureTest = false

static

Default flag for enabling different layers in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultNumHiddenLayersNum = 1

static

Default value for number of hidden layers in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultNumHiddenNeuronsFirstMin = 5

static

Default value for number of hidden neurons in first hidden layer in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultNumHiddenNeuronsFirstMax = 25

static

Default value for number of hidden neurons in first hidden layer in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultNumHiddenNeuronsFirstNum = 1

static

Default value for number of hidden neurons in first hidden layer in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultNumHiddenNeuronsSecondMin = 2

static

Default value for number of hidden neurons in second hidden layer in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultNumHiddenNeuronsSecondMax = 10

static

Default value for number of hidden neurons in second hidden layer in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultNumHiddenNeuronsSecondNum = 1

static

Default value for number of hidden neurons in second hidden layer in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultNumHiddenNeuronsThirdMin = 2

static

Default value for number of hidden neurons in third hidden layer in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultNumHiddenNeuronsThirdMax = 10

static

Default value for number of hidden neurons in third hidden layer in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultNumHiddenNeuronsThirdNum = 1

static

Default value for number of hidden neurons in third hidden layer in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultMaxEpochs = 1000

static

Default value for maximum number of epochs in neural networks.

int IG.Num.NeuralTrainingLimits.DefaultEpochBundle = 100

static

Default value for number of epochs in bundle in neural networks.

bool IG.Num.NeuralTrainingLimits.DefaultEnableRangeTolerance = false

static

Flag for enabling toelrance that represent a percentage of the output range.

int IG.Num.NeuralTrainingLimits.DefaultInputLenght = 5

static

Default value for number of input neurons.

int IG.Num.NeuralTrainingLimits.DefaultOutputLenght = 3

static

Default value for number of output neurons.

double IG.Num.NeuralTrainingLimits._learningRateMin = NeuralTrainingParameters.DefaultLearningRate

protected

double IG.Num.NeuralTrainingLimits._learningRateMax = NeuralTrainingParameters.DefaultLearningRate

protected

int IG.Num.NeuralTrainingLimits._learningRateNum = 1

protected

double IG.Num.NeuralTrainingLimits._momentumMin = DefaultMomentumMin

protected

double IG.Num.NeuralTrainingLimits._momentumMax = DefaultMomentumMax

protected

int IG.Num.NeuralTrainingLimits._momentumNum = DefaultMomentumNum

protected

double IG.Num.NeuralTrainingLimits._alphaMin = DefaultAlphaMin

protected

double IG.Num.NeuralTrainingLimits._alphaMax = DefaultAlphaMax

protected

int IG.Num.NeuralTrainingLimits._alphaNum = DefaultAlphaNum

protected

double IG.Num.NeuralTrainingLimits._inputSafetyFactorMin = DefaultInputSafetyFactorMin

protected

double IG.Num.NeuralTrainingLimits._inputSafetyFactorMax = DefaultInputSafetyFactorMax

protected

int IG.Num.NeuralTrainingLimits._inputSafetyFactorNum = DefaultInputSafetyFactorNum

protected

double IG.Num.NeuralTrainingLimits._outputSafetyFactorMin = DefaultOutputSafetyFactorMin

protected

double IG.Num.NeuralTrainingLimits._outputSafetyFactorMax = DefaultOutputSafetyFactorMax

protected

int IG.Num.NeuralTrainingLimits._outputSafetyFactorNum = DefaultOutputSafetyFactorNum

protected

int IG.Num.NeuralTrainingLimits._maxEpochs = DefaultMaxEpochs

protected

int IG.Num.NeuralTrainingLimits._epochBundle = DefaultEpochBundle

protected

int IG.Num.NeuralTrainingLimits._inputLenght = DefaultInputLenght

protected

int IG.Num.NeuralTrainingLimits._outputLenght = DefaultOutputLenght

protected

bool IG.Num.NeuralTrainingLimits._enableRangeTolerance = DefaultEnableRangeTolerance

protected

IVector IG.Num.NeuralTrainingLimits._toleranceRms

protected

IVector IG.Num.NeuralTrainingLimits._toleranceMax

protected

IVector IG.Num.NeuralTrainingLimits._tolRmsRelative

private

Auxiliary properties for defining tolerances in a relative way:

double IG.Num.NeuralTrainingLimits._tolRmsRelativeScalar = NeuralTrainingParameters.DefaultToleranceRmsRelativeToRangeScalar

private

IVector IG.Num.NeuralTrainingLimits._tolMaxRelative

private

double IG.Num.NeuralTrainingLimits._tolMaxRelativeScalar = NeuralTrainingParameters.DefaultToleranceMaxRelativeToRangeScalar

private

bool IG.Num.NeuralTrainingLimits._enableArchitectureTest = DefaultEnableArchitectureTest

protected

int IG.Num.NeuralTrainingLimits._numHiddenLayersNum = DefaultNumHiddenLayersNum

protected

int IG.Num.NeuralTrainingLimits._numHiddenNeuronsFirstMin = DefaultNumHiddenNeuronsFirstMin

protected

int IG.Num.NeuralTrainingLimits._numHiddenNeuronsFirstMax = DefaultNumHiddenNeuronsFirstMax

protected

int IG.Num.NeuralTrainingLimits._numHiddenNeuronsFirstNum = DefaultNumHiddenNeuronsFirstNum

protected

int [] IG.Num.NeuralTrainingLimits._numHiddenNeuronsFirstValues

protected

bool IG.Num.NeuralTrainingLimits.randomTableEnabled = false

protected

int IG.Num.NeuralTrainingLimits._numHiddenNeuronsSecondMin = DefaultNumHiddenNeuronsSecondMin

protected

int IG.Num.NeuralTrainingLimits._numHiddenNeuronsSecondMax = DefaultNumHiddenNeuronsSecondMax

protected

int IG.Num.NeuralTrainingLimits._numHiddenNeuronsSecondNum = DefaultNumHiddenNeuronsSecondNum

protected

int [] IG.Num.NeuralTrainingLimits._numHiddenNeuronsSecondValues

protected

int IG.Num.NeuralTrainingLimits._numHiddenNeuronsThirdMin = DefaultNumHiddenNeuronsThirdMin

protected

int IG.Num.NeuralTrainingLimits._numHiddenNeuronsThirdMax = DefaultNumHiddenNeuronsThirdMax

protected

int IG.Num.NeuralTrainingLimits._numHiddenNeuronsThirdNum = DefaultNumHiddenNeuronsThirdNum

protected

int [] IG.Num.NeuralTrainingLimits._numHiddenNeuronsThirdValues

protected

Property Documentation

double IG.Num.NeuralTrainingLimits.LearningRateMin

getset

Minimum limit for learning rate.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

double IG.Num.NeuralTrainingLimits.LearningRateMax

getset

Maximum limit for learning rate.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.LearningRateNum

getset

Number of learning rates.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

double IG.Num.NeuralTrainingLimits.MomentumMin

getset

Minimum limit for momentum.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

double IG.Num.NeuralTrainingLimits.MomentumMax

getset

Maximum limit for momentum.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.MomentumNum

getset

Number of momentums.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

double IG.Num.NeuralTrainingLimits.AlphaMin

getset

Minimum limit for alpha value.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

double IG.Num.NeuralTrainingLimits.AlphaMax

getset

Maximum limit for alpha value.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.AlphaNum

getset

Number of alpha value.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

double IG.Num.NeuralTrainingLimits.InputSafetyFactorMin

getset

Minimum limit for input safety factor value.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

double IG.Num.NeuralTrainingLimits.InputSafetyFactorMax

getset

Maximum limit for input safety factor value.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.InputSafetyFactorNum

getset

Number of input safety factor values.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

double IG.Num.NeuralTrainingLimits.OutputSafetyFactorMin

getset

Minimum limit for output safety factor value.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

double IG.Num.NeuralTrainingLimits.OutputSafetyFactorMax

getset

Maximum limit for output safety factor value.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.OutputSafetyFactorNum

getset

Number of output safety factor values.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.MaxEpochs

getset

Maximum number of epochs performed in training.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.EpochBundle

getset

Number of epochs in boundle.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.InputLenght

getset

Number of input neurons.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.OutputLength

getset

Number of output neurons.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

bool IG.Num.NeuralTrainingLimits.EnableRangeTolerance

getset

Flag for enabling toelrance that represent a percentage of the output range.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

IVector IG.Num.NeuralTrainingLimits.ToleranceRms

getset

Tolerance for RMS.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

IVector IG.Num.NeuralTrainingLimits.ToleranceMax

getset

Maximum tolerance for max. abs. difference.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

IVector IG.Num.NeuralTrainingLimits.ToleranceRmsRelativeToRange

getset

Relative tolerances on RMS errors of outputs over training points, relative to the correspoinding ranges of output data.

See also: NeuralApproximatorBase.ToleranceRmsRelativeToRange

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

double IG.Num.NeuralTrainingLimits.ToleranceRmsRelativeToRangeScalar

getset

Scalar through which all components of the Relative tolerances on RMS errors of outputs can be set to the same value.

See also: NeuralApproximatorBase.ToleranceRmsRelativeToRangeScalar

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

IVector IG.Num.NeuralTrainingLimits.ToleranceMaxRelativeToRange

getset

Relative tolerances on max. abs. errors of outputs over training points, relative to the correspoinding ranges of output data.

See also: NeuralApproximatorBase.ToleranceMaxRelativeToRange

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

double IG.Num.NeuralTrainingLimits.ToleranceMaxRelativeToRangeScalar

getset

Scalar through which all components of the Relative tolerances on max. abs. errors of outputs can be set to the same value.

See also: NeuralApproximatorBase.ToleranceMaxRelativeToRangeScalar

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

bool IG.Num.NeuralTrainingLimits.EnableArchitectureTest

getset

Flag for enabling test in architecture of ANN.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.NumHiddenLayersNum

getset

Number of hidden layers in neural network.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.NumHiddenNeuronsFirstMin

getset

Minimum number of hidden neurons in first hidden layer.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.NumHiddenNeuronsFirstMax

getset

Maximum number of hidden neurons in first hidden layer.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.NumHiddenNeuronsFirstNum

getset

Number of hidden neurons in first hidden layer.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int [] IG.Num.NeuralTrainingLimits.NumHiddenNeuronsFirstValues

getset

Values of number of hidden neurons in the first layers that will appear in the table.

Specifying an array of values for the number of hidden neurons in the first layer makes possible to use custom values, not just equidistant values between minimal and maximal value.

If array of values is set, the number of values is set to the length of the array, minimal value is set to its first element and maximal value is set to its last element.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.NumHiddenNeuronsSecondMin

getset

Minimum number of hidden neurons in second hidden layer.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.NumHiddenNeuronsSecondMax

getset

Maximum number of hidden neurons in second hidden layer.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.NumHiddenNeuronsSecondNum

getset

Number of hidden neurons in second hidden layer.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int [] IG.Num.NeuralTrainingLimits.NumHiddenNeuronsSecondValues

getset

Values of number of hidden neurons in the second layers that will appear in the table.

Specifying an array of values for the number of hidden neurons in the second layer makes possible to use custom values, not just equidistant values between minimal and maximal value.

If array of values is set, the number of values is set to the length of the array, minimal value is set to its first element and maximal value is set to its last element.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.NumHiddenNeuronsThirdMin

getset

Minimum number of hidden neurons in third hidden layer.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.NumHiddenNeuronsThirdMax

getset

Maximum number of hidden neurons in third hidden layer.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int IG.Num.NeuralTrainingLimits.NumHiddenNeuronsThirdNum

getset

Number of hidden neurons in third hidden layer.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

int [] IG.Num.NeuralTrainingLimits.NumHiddenNeuronsThirdValues

getset

Values of number of hidden neurons in the third layer that will appear in the table.

Specifying an array of values for the number of hidden neurons in the third layer makes possible to use custom values, not just equidistant values between minimal and maximal value.

If the array of values is set, the number of values is set to the length of the array, minimal value is set to its first element and maximal value is set to its last element.

Referenced by IG.Num.NeuralTrainingLimitsDto.CopyFromPlain().

The documentation for this class was generated from the following file:

igbase/numeric/approximation/neural/NeuralTrainingParameters.cs

Public Member Functions

Static Public Member Functions

Static Public Attributes

Protected Member Functions

Protected Attributes

Properties

Private Member Functions

Private Attributes

Detailed Description

Member Function Documentation

Member Data Documentation

Property Documentation