C# Class Encog.Normalize.DataNormalization

Inheritance: Encog.Persist.BasicPersistedObject

Afficher le fichier Open project: encog/encog-silverlight-core

Méthodes publiques

Méthode	Description
AddInputField ( IInputField f ) : void	Add an input field.
AddOutputField ( IOutputField outputField ) : void	Add an output field. This output field will be added as a "neural network input field", not an "ideal output field".
AddOutputField ( IOutputField outputField, bool ideal ) : void	Add a field and allow it to be specified as an "ideal output field". An "ideal" field is the expected output that the neural network is training towards.
AddSegregator ( ISegregator segregator ) : void	Add a segregator.
BuildForNetworkInput ( double data ) : INeuralData	Build "input data for a neural network" based on the input values provided. This allows input for a neural network to be normalized. This is typically used when data is to be presented to a trained neural network.
Clone ( ) : object	Not implemented.
CreatePersistor ( ) : IPersistor	A persistor to persist this DataNormalization object.
FindInputField ( Type clazz, int count ) : IInputField	Find an input field by its class.
FindOutputField ( Type clazz, int count ) : IOutputField	Find an output field by its class.
GetNetworkInputLayerSize ( ) : int	Calculate the number of output fields that are not used as ideal values, these will be the input to the neural network. This is the input layer size for the neural network.
GetNetworkOutputLayerSize ( ) : int	The number of output fields that are used as ideal values, these will be the ideal output from the neural network. This is the output layer size for the neural network.
GetOutputFieldCount ( ) : int	The total size of all output fields. This takes into account output fields that generate more than one value.
InitForOutput ( ) : void	Setup the row for output.
InitForPass ( ) : void	Setup the row for output.
Process ( ) : void	Call this method to begin the normalization process. Any status updates will be sent to the class specified in the constructor.
TwoPassesNeeded ( ) : bool	Determine if two passes will be needed.

Private Methods

Méthode	Description
ApplyMinMax ( ) : void	Called internally to allow each of the input fields to update their min/max values in the first pass.
DetermineInputFieldValue ( IInputField field, int index ) : double	Called internally to obtain the current value for an input field.
DetermineInputFieldValues ( int index ) : void	Called internally to determine all of the input field values.
FirstPass ( ) : void	First pass, count everything, establish min/max.
Next ( ) : bool	Called internally to advance to the next row.
OpenCSV ( ) : void	Called internally to open the CSV file.
OpenDataSet ( ) : void	Open any datasets that were used by the input layer.
ReportResult ( String message, int total, int current ) : void	Report on the current progress.
SecondPass ( ) : void	The second pass actually writes the data to the output files.
ShouldInclude ( ) : bool	Should this row be included? Check the segregatprs.

Method Details

AddInputField() public méthode

Add an input field.

public AddInputField ( IInputField f ) : void
f	IInputField	The input field to add.
Résultat	void

AddOutputField() public méthode

Add an output field. This output field will be added as a "neural network input field", not an "ideal output field".

public AddOutputField ( IOutputField outputField ) : void
outputField	IOutputField	The output field to add.
Résultat	void

AddOutputField() public méthode

Add a field and allow it to be specified as an "ideal output field". An "ideal" field is the expected output that the neural network is training towards.

public AddOutputField ( IOutputField outputField, bool ideal ) : void
outputField	IOutputField	The output field.
ideal	bool	True if this is an ideal field.
Résultat	void

AddSegregator() public méthode

Add a segregator.

public AddSegregator ( ISegregator segregator ) : void
segregator	ISegregator	The segregator to add.
Résultat	void

BuildForNetworkInput() public méthode

Build "input data for a neural network" based on the input values provided. This allows input for a neural network to be normalized. This is typically used when data is to be presented to a trained neural network.

public BuildForNetworkInput ( double data ) : INeuralData
data	double	The input values to be normalized.
Résultat	INeuralData

Clone() public méthode

Not implemented.

public Clone ( ) : object
Résultat	object

CreatePersistor() public méthode

A persistor to persist this DataNormalization object.

public CreatePersistor ( ) : IPersistor
Résultat	IPersistor

FindInputField() public méthode

Find an input field by its class.

public FindInputField ( Type clazz, int count ) : IInputField
clazz	System.Type	The input field class type you are looking for.
count	int	The instance of the input field needed, 0 for the first.
Résultat	IInputField

FindOutputField() public méthode

Find an output field by its class.

public FindOutputField ( Type clazz, int count ) : IOutputField
clazz	System.Type	The output field class type you are looking for.
count	int	The instance of the output field needed, 0 for the first.
Résultat	IOutputField

GetNetworkInputLayerSize() public méthode

Calculate the number of output fields that are not used as ideal values, these will be the input to the neural network. This is the input layer size for the neural network.

public GetNetworkInputLayerSize ( ) : int
Résultat	int

GetNetworkOutputLayerSize() public méthode

The number of output fields that are used as ideal values, these will be the ideal output from the neural network. This is the output layer size for the neural network.

public GetNetworkOutputLayerSize ( ) : int
Résultat	int

GetOutputFieldCount() public méthode

The total size of all output fields. This takes into account output fields that generate more than one value.

public GetOutputFieldCount ( ) : int
Résultat	int

InitForOutput() public méthode

Setup the row for output.

public InitForOutput ( ) : void
Résultat	void

InitForPass() public méthode

Setup the row for output.

public InitForPass ( ) : void
Résultat	void

Process() public méthode

Call this method to begin the normalization process. Any status updates will be sent to the class specified in the constructor.

public Process ( ) : void
Résultat	void

TwoPassesNeeded() public méthode

Determine if two passes will be needed.

public TwoPassesNeeded ( ) : bool
Résultat	bool