jsat.datatransform.featureselection

Class MutualInfoFS

java.lang.Object

jsat.datatransform.RemoveAttributeTransform

jsat.datatransform.featureselection.MutualInfoFS

All Implemented Interfaces:

Serializable, Cloneable, DataTransform

public class MutualInfoFS
extends RemoveAttributeTransform

Performs greedy feature selection based on Mutual Information of the features with respect to the class values. This is an attempt to select features that are discriminative for classification tasks.
The method of performing Mutual Information on numeric attributes is controlled by MutualInfoFS.NumericalHandeling

Author:

Edward Raff

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	MutualInfoFS.NumericalHandeling The definition for mutual information for continuous attributes requires an integration of an unknown function, as such requires some form of approximation.

Field Summary

Fields inherited from class jsat.datatransform.RemoveAttributeTransform

catIndexMap, numIndexMap

Constructor Summary

Constructors

Modifier	Constructor and Description
	MutualInfoFS() Creates a new Mutual Information feature selection object that attempts to select up to 100 features.
	MutualInfoFS(ClassificationDataSet dataSet, int featureCount) Creates a new Mutual Information feature selection object.
	MutualInfoFS(ClassificationDataSet dataSet, int featureCount, MutualInfoFS.NumericalHandeling numericHandling) Creates a new Mutual Information feature selection object.
	MutualInfoFS(int featureCount) Creates a new Mutual Information feature selection object.
	MutualInfoFS(int featureCount, MutualInfoFS.NumericalHandeling numericHandling) Creates a new Mutual Information feature selection object.
protected	MutualInfoFS(MutualInfoFS toCopy) Copy constructor

Method Summary

Modifier and Type	Method and Description
MutualInfoFS	clone()
void	fit(DataSet data) Fits this transform to the given dataset.
int	getFeatureCount() Returns the number of features to select
MutualInfoFS.NumericalHandeling	getHandling() Returns the method of numericHandling numeric features
void	setFeatureCount(int featureCount) Sets the number of features to select
void	setHandling(MutualInfoFS.NumericalHandeling handling) Sets the method of numericHandling numeric features

Methods inherited from class jsat.datatransform.RemoveAttributeTransform

consolidate, getKeptNominal, getKeptNumeric, getReverseNominalMap, getReverseNumericMap, setUp, transform

Methods inherited from class java.lang.Object

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

Constructor Detail

MutualInfoFS

public MutualInfoFS()

Creates a new Mutual Information feature selection object that attempts to select up to 100 features. Numeric attributes are handled by MutualInfoFS.NumericalHandeling.BINARY

MutualInfoFS

public MutualInfoFS(int featureCount)

Creates a new Mutual Information feature selection object. Numeric attributes are handled by MutualInfoFS.NumericalHandeling.BINARY

Parameters:

featureCount - the number of features to select

MutualInfoFS

public MutualInfoFS(ClassificationDataSet dataSet,
                    int featureCount)

Creates a new Mutual Information feature selection object. Numeric attributes are handled by MutualInfoFS.NumericalHandeling.BINARY

Parameters:

dataSet - the classification data set to perform feature selection from

featureCount - the number of features to select

MutualInfoFS

protected MutualInfoFS(MutualInfoFS toCopy)

Copy constructor

Parameters:

toCopy - the object to copy

MutualInfoFS

public MutualInfoFS(int featureCount,
                    MutualInfoFS.NumericalHandeling numericHandling)

Creates a new Mutual Information feature selection object.

Parameters:

featureCount - the number of features to select

numericHandling - the way to handle the computation of mutual information for numeric attributes

MutualInfoFS

public MutualInfoFS(ClassificationDataSet dataSet,
                    int featureCount,
                    MutualInfoFS.NumericalHandeling numericHandling)

Creates a new Mutual Information feature selection object.

Parameters:

dataSet - the classification data set to perform feature selection from

featureCount - the number of features to select

numericHandling - the way to handle the computation of mutual information for numeric attributes

Method Detail

fit

public void fit(DataSet data)

Description copied from interface: DataTransform

Fits this transform to the given dataset. Some transforms can only be learned from classification or regression datasets. If an incompatible dataset type is given, a FailedToFitException exception may be thrown.

Specified by:

fit in interface DataTransform

Overrides:

fit in class RemoveAttributeTransform

Parameters:

data - the dataset to fir this transform to

clone

public MutualInfoFS clone()

Specified by:

clone in interface DataTransform

Overrides:

clone in class RemoveAttributeTransform

setFeatureCount

public void setFeatureCount(int featureCount)

Sets the number of features to select

Parameters:

featureCount - the number of features to select

getFeatureCount

public int getFeatureCount()

Returns the number of features to select

Returns:

the number of features to select

setHandling

public void setHandling(MutualInfoFS.NumericalHandeling handling)

Sets the method of numericHandling numeric features

Parameters:

handling - the numeric numericHandling

getHandling

public MutualInfoFS.NumericalHandeling getHandling()

Returns the method of numericHandling numeric features

Returns:

the method of numericHandling numeric features