mulan.classifier.neural
Class NormalizationFilter

java.lang.Object
  mulan.classifier.neural.NormalizationFilter

All Implemented Interfaces:

Serializable

public class NormalizationFilter
extends Object
implements Serializable

Performs a normalization of numeric attributes of the data set. It is initialized based on given MultiLabelInstances data set and then can be used to normalize Instance instances which conform to the format of the data set the NormalizationFilter was initialized from.

Version:

2012.02.27

Author:

Jozef Vilcek

See Also:

Serialized Form

Constructor Summary

NormalizationFilter(MultiLabelInstances mlData, boolean performNormalization)
Creates a new instance of NormalizationFilter class for given data set.

NormalizationFilter(MultiLabelInstances mlData, boolean performNormalization, double minValue, double maxValue)
Creates a new instance of NormalizationFilter class for given data set.

Method Summary

void	normalize(Instance instance) Performs a normalization of numerical attributes on given instance.

Methods inherited from class java.lang.Object

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

Constructor Detail

NormalizationFilter

public NormalizationFilter(MultiLabelInstances mlData,
                           boolean performNormalization,
                           double minValue,
                           double maxValue)

Creates a new instance of NormalizationFilter class for given data set.

Parameters:

mlData - the MultiLabelInstances data set from which normalizer should be initialized.

performNormalization - indicates whether normalization of instances contained in the data set used for initialization should be performed

minValue - the minimum value of the normalization range for numerical attributes

maxValue - the maximum value of the normalization range for numerical attributes

NormalizationFilter

public NormalizationFilter(MultiLabelInstances mlData,
                           boolean performNormalization)

Creates a new instance of NormalizationFilter class for given data set. The normalizer will be initialized to perform normalization to the default range <-1,1>.

Parameters:

mlData - the MultiLabelInstances data set from which normalizer should be initialized.

performNormalization - indicates whether normalization of instances contained in the data set used for initialization should be performed

Method Detail

normalize

public void normalize(Instance instance)

Performs a normalization of numerical attributes on given instance. The instance must conform to format of instances data the NormalizationFilter was initialized with.

Parameters:

instance - the instance to be normalized