All Packages  Class Hierarchy  This Package  Previous  Next  Index  WEKA's home

Class weka.attributeSelection.WrapperSubsetEval

java.lang.Object
   |
   +----weka.attributeSelection.ASEvaluation
           |
           +----weka.attributeSelection.SubsetEvaluator
                   |
                   +----weka.attributeSelection.WrapperSubsetEval

public class WrapperSubsetEval

extends SubsetEvaluator

implements OptionHandler

For more information see:
Kohavi, R., John G., Wrappers for Feature Subset Selection. In Artificial Intelligence journal, special issue on relevance, Vol. 97, Nos 1-2, pp.273-324.

Valid options are:

-B
Class name of base learner to use for accuracy estimation. Place any classifier options last on the command line following a "--". Eg -B weka.classifiers.NaiveBayes ... -- -K

-F
Number of cross validation folds to use for estimating accuracy.

-T
Threshold by which to execute another cross validation (standard deviation ---expressed as a percentage of the mean).

Version:

$Revision: 1.8 $

Author:

Mark Hall (mhall@cs.waikato.ac.nz)

WrapperSubsetEval()

Constructor.

buildEvaluator(Instances)

Generates a attribute evaluator.

evaluateSubset(BitSet)

Evaluates a subset of attributes

getClassifier()

Get the classifier used as the base learner.

getFolds()

Get the number of folds used for accuracy estimation

getOptions()

Gets the current settings of WrapperSubsetEval.

getSeed()

Get the random number seed used for cross validation

getThreshold()

Get the value of the threshold

listOptions()

Returns an enumeration describing the available options

main(String[])

Main method for testing this class.

setClassifier(Classifier)

Set the classifier to use for accuracy estimation

setFolds(int)

Set the number of folds to use for accuracy estimation

setOptions(String[])

Parses a given list of options.

setSeed(int)

Set the seed to use for cross validation

setThreshold(double)

Set the value of the threshold for repeating cross validation

toString()

Returns a string describing the wrapper

 public WrapperSubsetEval()

Constructor. Calls restOptions to set default options

 public Enumeration listOptions()

Returns an enumeration describing the available options

Returns:

an enumeration of all the available options

 public void setOptions(String options[]) throws Exception

Parses a given list of options. Valid options are:

-B
Class name of base learner to use for accuracy estimation. Place any classifier options last on the command line following a "--". Eg -B weka.classifiers.NaiveBayes ... -- -K

-F
Number of cross validation folds to use for estimating accuracy.

-T
Threshold by which to execute another cross validation (standard deviation ---expressed as a percentage of the mean).

Parameters:

options - the list of options as an array of strings

Throws: Exception

if an option is not supported

 public void setThreshold(double t)

Set the value of the threshold for repeating cross validation

Parameters:

t - the value of the threshold

 public double getThreshold()

Get the value of the threshold

Returns:

the threshold as a double

 public void setFolds(int f)

Set the number of folds to use for accuracy estimation

Parameters:

f - the number of folds

 public int getFolds()

Get the number of folds used for accuracy estimation

Returns:

the number of folds

 public void setSeed(int s)

Set the seed to use for cross validation

Parameters:

s - the seed

 public int getSeed()

Get the random number seed used for cross validation

Returns:

the seed

 public void setClassifier(Classifier newClassifier)

Set the classifier to use for accuracy estimation

Parameters:

newClassifier - the Classifier to use.

 public Classifier getClassifier()

Get the classifier used as the base learner.

Returns:

the classifier used as the classifier

 public String[] getOptions()

Gets the current settings of WrapperSubsetEval.

Returns:

an array of strings suitable for passing to setOptions()

 public void buildEvaluator(Instances data) throws Exception

Generates a attribute evaluator. Has to initialize all fields of the evaluator that are not being set via options.

Parameters:

data - set of instances serving as training data

Throws: Exception

if the evaluator has not been generated successfully

Overrides:

buildEvaluator in class ASEvaluation

 public double evaluateSubset(BitSet subset) throws Exception

Evaluates a subset of attributes

Parameters:

subset - a bitset representing the attribute subset to be evaluated

Throws: Exception

if the subset could not be evaluated

Overrides:

evaluateSubset in class SubsetEvaluator

 public String toString()

Returns a string describing the wrapper

Returns:

the description as a string

Overrides:

toString in class Object

 public static void main(String args[])

Main method for testing this class.

Parameters:

args - the options

All Packages  Class Hierarchy  This Package  Previous  Next  Index  WEKA's home