org.encog.ml.data.versatile.normalizers.strategies

Class BasicNormalizationStrategy

java.lang.Object

org.encog.ml.data.versatile.normalizers.strategies.BasicNormalizationStrategy

All Implemented Interfaces:

NormalizationStrategy

public class BasicNormalizationStrategy
extends Object
implements NormalizationStrategy

Provides a basic normalization strategy that will work with most models built into Encog. This is often used as a starting point for building more customized models, as this normalizer works mainly by using maps to define which normalizer to use for what data type.

Constructor Summary

Constructors

Constructor and Description
BasicNormalizationStrategy() Default constructor.
BasicNormalizationStrategy(double inputLow, double inputHigh, double outputLow, double outputHigh) Construct the basic normalization strategy.

Method Summary

Methods

Modifier and Type	Method and Description
void	assignInputNormalizer(ColumnType colType, Normalizer norm) Assign a normalizer to the specified column type for output.
void	assignOutputNormalizer(ColumnType colType, Normalizer norm) Assign a normalizer to the specified column type for output.
String	denormalizeColumn(ColumnDefinition colDef, boolean isInput, MLData data, int dataColumn) Normalize a column, with a double input.
Map<ColumnType,Normalizer>	getInputNormalizers()
Map<ColumnType,Normalizer>	getOutputNormalizers()
int	normalizeColumn(ColumnDefinition colDef, boolean isInput, double value, double[] outputData, int outputColumn) Normalize a column, with a double value.
int	normalizeColumn(ColumnDefinition colDef, boolean isInput, String value, double[] outputData, int outputColumn) Normalize a column, with a string input.
int	normalizedSize(ColumnDefinition colDef, boolean isInput) Calculate how many elements a column will normalize into.

Methods inherited from class java.lang.Object

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

Constructor Detail

BasicNormalizationStrategy

public BasicNormalizationStrategy(double inputLow,
                          double inputHigh,
                          double outputLow,
                          double outputHigh)

Construct the basic normalization strategy.

Parameters:

inputLow - The desired low to normalize input into.

inputHigh - The desired high to normalize input into.

outputLow - The desired low to normalize output into.

outputHigh - The desired high to normalize output into.

BasicNormalizationStrategy

public BasicNormalizationStrategy()

Default constructor.

Method Detail

assignInputNormalizer

public void assignInputNormalizer(ColumnType colType,
                         Normalizer norm)

Assign a normalizer to the specified column type for output.

Parameters:

colType - The column type.

norm - The normalizer.

assignOutputNormalizer

public void assignOutputNormalizer(ColumnType colType,
                          Normalizer norm)

Assign a normalizer to the specified column type for output.

Parameters:

colType - The column type.

norm - The normalizer.

normalizedSize

public int normalizedSize(ColumnDefinition colDef,
                 boolean isInput)

Calculate how many elements a column will normalize into.

Specified by:

normalizedSize in interface NormalizationStrategy

Parameters:

colDef - The column definition.

isInput - True, if this is an input column.

Returns:

The number of elements needed to normalize this column.

normalizeColumn

public int normalizeColumn(ColumnDefinition colDef,
                  boolean isInput,
                  String value,
                  double[] outputData,
                  int outputColumn)

Normalize a column, with a string input.

Specified by:

normalizeColumn in interface NormalizationStrategy

Parameters:

colDef - The column definition.

isInput - True, if this is an input column.

value - The value to normalize.

outputData - The output data.

outputColumn - The element to begin outputing to.

Returns:

The new output element, advanced by the correct amount.

normalizeColumn

public int normalizeColumn(ColumnDefinition colDef,
                  boolean isInput,
                  double value,
                  double[] outputData,
                  int outputColumn)

Normalize a column, with a double value.

Specified by:

normalizeColumn in interface NormalizationStrategy

Parameters:

colDef - The column definition.

isInput - True, if this is an input column.

value - The value to normalize.

outputData - The output data.

outputColumn - The element to begin outputing to.

Returns:

The new output element, advanced by the correct amount.

getInputNormalizers

public Map<ColumnType,Normalizer> getInputNormalizers()

Returns:

the inputNormalizers

getOutputNormalizers

public Map<ColumnType,Normalizer> getOutputNormalizers()

Returns:

the outputNormalizers

denormalizeColumn

public String denormalizeColumn(ColumnDefinition colDef,
                       boolean isInput,
                       MLData data,
                       int dataColumn)

Normalize a column, with a double input.

Specified by:

denormalizeColumn in interface NormalizationStrategy

Parameters:

colDef - The column definition.

isInput - True, if this is an input column.

Returns:

The new output element, advanced by the correct amount.