jsat.datatransform.featureselection

Class BDS

java.lang.Object

jsat.datatransform.featureselection.BDS

All Implemented Interfaces:

Serializable, Cloneable, DataTransform

public class BDS
extends Object
implements DataTransform

Bidirectional Search (BDS) is a greedy method of selecting a subset of features to use for prediction. It performs both SFS and SBS search at the same time. At each step, a feature is greedily added to one set, and then a feature greedily removed from another set. Once a feature is added / removed in one set, it is unavailable for selection in the other. This can be used to select up to half of the original features.

Author:

Edward Raff

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
BDS(BDS toClone) Copy constructor
BDS(int featureCount, ClassificationDataSet dataSet, Classifier evaluator, int folds) Performs BDS feature selection for a classification problem
BDS(int featureCount, Classifier evaluator, int folds) Creates a BDS feature selection for a classification problem
BDS(int featureCount, RegressionDataSet dataSet, Regressor evaluator, int folds) Performs BDS feature selection for a regression problem
BDS(int featureCount, Regressor evaluator, int folds) Creates a BDS feature selection for a regression problem

Method Summary

Modifier and Type	Method and Description
BDS	clone()
void	fit(DataSet data) Fits this transform to the given dataset.
int	getFeatureCount() Returns the number of features to use
int	getFolds()
Set<Integer>	getSelectedCategorical() Returns a copy of the set of categorical features selected by the search algorithm
Set<Integer>	getSelectedNumerical() Returns a copy of the set of numerical features selected by the search algorithm.
void	setFeatureCount(int featureCount) Sets the number of features to select for use from the set of all input features
void	setFolds(int folds) Sets the number of folds to use for cross validation when estimating the error rate
DataPoint	transform(DataPoint dp) Returns a new data point that is a transformation of the original data point.

Methods inherited from class java.lang.Object

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

Constructor Detail

BDS

public BDS(BDS toClone)

Copy constructor

Parameters:

toClone -

BDS

public BDS(int featureCount,
           Classifier evaluator,
           int folds)

Creates a BDS feature selection for a classification problem

Parameters:

featureCount - the number of features to select

dataSet - the data set to perform feature selection on

evaluator - the classifier to use in determining accuracy given a feature subset

folds - the number of cross validation folds to use in selection

BDS

public BDS(int featureCount,
           ClassificationDataSet dataSet,
           Classifier evaluator,
           int folds)

Performs BDS feature selection for a classification problem

Parameters:

featureCount - the number of features to select

dataSet - the data set to perform feature selection on

evaluator - the classifier to use in determining accuracy given a feature subset

folds - the number of cross validation folds to use in selection

BDS

public BDS(int featureCount,
           Regressor evaluator,
           int folds)

Creates a BDS feature selection for a regression problem

Parameters:

featureCount - the number of features to select

evaluator - the regressor to use in determining accuracy given a feature subset

folds - the number of cross validation folds to use in selection

BDS

public BDS(int featureCount,
           RegressionDataSet dataSet,
           Regressor evaluator,
           int folds)

Performs BDS feature selection for a regression problem

Parameters:

featureCount - the number of features to select

dataSet - the data set to perform feature selection on

evaluator - the regressor to use in determining accuracy given a feature subset

folds - the number of cross validation folds to use in selection

Method Detail

transform

public DataPoint transform(DataPoint dp)

Description copied from interface: DataTransform

Returns a new data point that is a transformation of the original data point. This new data point is a different object, but may contain the same references as the original data point. It is not guaranteed that you can mutate the transformed point without having a side effect on the original point.

Specified by:

transform in interface DataTransform

Parameters:

dp - the data point to apply a transformation to

Returns:

a transformed data point

clone

public BDS clone()

Specified by:

clone in interface DataTransform

Overrides:

clone in class Object

getSelectedCategorical

public Set<Integer> getSelectedCategorical()

Returns a copy of the set of categorical features selected by the search algorithm

Returns:

the set of categorical features to use

getSelectedNumerical

public Set<Integer> getSelectedNumerical()

Returns a copy of the set of numerical features selected by the search algorithm.

Returns:

the set of numeric features to use

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

Parameters:

data - the dataset to fir this transform to

setFeatureCount

public void setFeatureCount(int featureCount)

Sets the number of features to select for use from the set of all input features

Parameters:

featureCount - the number of features to use

getFeatureCount

public int getFeatureCount()

Returns the number of features to use

Returns:

the number of features to use

setFolds

public void setFolds(int folds)

Sets the number of folds to use for cross validation when estimating the error rate

Parameters:

folds - the number of folds to use for cross validation when estimating the error rate

getFolds

public int getFolds()

Returns:

the number of folds to use for cross validation when estimating the error rate