SVMAttributeEval (Documentation for extended WEKA including Ensembles of Hierarchically Nested Dichotomies)

Overview

Package

Class

Use

Tree

Deprecated

Index

Help

PREV CLASS NEXT CLASS

FRAMES NO FRAMES

SUMMARY: NESTED | FIELD | CONSTR | METHOD

DETAIL: FIELD | CONSTR | METHOD

weka.attributeSelection
Class SVMAttributeEval

java.lang.Object
  weka.attributeSelection.ASEvaluation
      weka.attributeSelection.AttributeEvaluator
          weka.attributeSelection.SVMAttributeEval

All Implemented Interfaces:: OptionHandler, java.io.Serializable

public class SVMAttributeEval
extends AttributeEvaluator
implements OptionHandler

Class for Evaluating attributes individually by using the SVM classifier. Attributes are ranked by the square of the weight assigned by the SVM. Attribute selection for multiclass problems is handled by ranking attributes for each class seperately using a one-vs-all method and then "dealing" from the top of each pile to give a final ranking.

Valid options are:

-X
Specify constant rate at which attributes are eliminated per invocation of the support vector machine. Default = 1.

-Y
Specify the percentage rate at which attributes are eliminated per invocation of the support vector machine. This setting trumps the constant rate setting. Default = 0 (percentage rate ignored).

-Z
Specify the threshold below which the percentage elimination method reverts to the constant elimination method.

-C
Specify the value of C - the complexity parameter to be passed on to the support vector machine.

-P
Sets the epsilon for round-off error. (default 1.0e-25)

-T
Sets the tolerance parameter. (default 1.0e-10)

Version:: $Revision: 1.17 $
Author:: Eibe Frank (eibe@cs.waikato.ac.nz), Mark Hall (mhall@cs.waikato.ac.nz)
See Also:: Serialized Form

Field Summary
`private double[]`	`m_attScores` The attribute scores
`private int`	`m_numToEliminate` Constant rate of attribute elimination per iteration
`private int`	`m_percentThreshold` Threshold below which percent elimination switches to constant elimination
`private int`	`m_percentToEliminate` Percentage rate of attribute elimination, trumps constant rate (above threshold), ignored if = 0
`private double`	`m_smoCParameter` Complexity parameter to pass on to SMO
`private int`	`m_smoFilterType` Filter parameter to pass on to SMO
`private double`	`m_smoPParameter` Epsilon parameter to pass on to SMO
`private double`	`m_smoTParameter` Tolerance parameter to pass on to SMO

Constructor Summary
`SVMAttributeEval()` Constructor

Method Summary
`java.lang.String`	`attsToEliminatePerIterationTipText()` Returns a tip text for this property suitable for display in the GUI
`void`	`buildEvaluator(Instances data)` Initializes the evaluator.
`java.lang.String`	`complexityParameterTipText()` Returns a tip text for this property suitable for display in the GUI
`java.lang.String`	`epsilonParameterTipText()` Returns a tip text for this property suitable for display in the GUI
`double`	`evaluateAttribute(int attribute)` Evaluates an attribute by returning the rank of the square of its coefficient in a linear support vector machine.
`java.lang.String`	`filterTypeTipText()` Returns a tip text for this property suitable for display in the GUI
`int`	`getAttsToEliminatePerIteration()` Get the constant rate of attribute elimination per iteration
`double`	`getComplexityParameter()` Get the value of C used with SMO
`double`	`getEpsilonParameter()` Get the value of P used with SMO
`SelectedTag`	`getFilterType()` Get the filtering mode passed to SMO
`java.lang.String[]`	`getOptions()` Gets the current settings of SVMAttributeEval
`int`	`getPercentThreshold()` Get the threshold below which percentage elimination reverts to constant elimination.
`int`	`getPercentToEliminatePerIteration()` Get the percentage rate of attribute elimination per iteration
`double`	`getToleranceParameter()` Get the value of T used with SMO
`java.lang.String`	`globalInfo()` Returns a string describing this attribute evaluator
`java.util.Enumeration`	`listOptions()` Returns an enumeration describing all the available options
`static void`	`main(java.lang.String[] args)` Main method for testing this class.
`java.lang.String`	`percentThresholdTipText()` Returns a tip text for this property suitable for display in the GUI
`java.lang.String`	`percentToEliminatePerIterationTipText()` Returns a tip text for this property suitable for display in the GUI
`private int[]`	`rankBySVM(int classInd, Instances data)` Get SVM-ranked attribute indexes (best to worst) selected for the class attribute indexed by classInd (one-vs-all).
`protected void`	`resetOptions()` Resets options to defaults.
`void`	`setAttsToEliminatePerIteration(int cRate)` Set the constant rate of attribute elimination per iteration
`void`	`setComplexityParameter(double svmC)` Set the value of C for SMO
`void`	`setEpsilonParameter(double svmP)` Set the value of P for SMO
`void`	`setFilterType(SelectedTag newType)` The filtering mode to pass to SMO
`void`	`setOptions(java.lang.String[] options)` Parses a given list of options Valid options are: -X Specify constant rate at which attributes are eliminated per invocation of the support vector machine.
`void`	`setPercentThreshold(int pThresh)` Set the threshold below which percentage elimination reverts to constant elimination.
`void`	`setPercentToEliminatePerIteration(int pRate)` Set the percentage of attributes to eliminate per iteration
`void`	`setToleranceParameter(double svmT)` Set the value of T for SMO
`java.lang.String`	`toleranceParameterTipText()` Returns a tip text for this property suitable for display in the GUI
`java.lang.String`	`toString()` Return a description of the evaluator

Methods inherited from class weka.attributeSelection.ASEvaluation

forName, makeCopies, postProcess

Methods inherited from class java.lang.Object

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

Field Detail

m_attScores

private double[] m_attScores

The attribute scores

m_numToEliminate

private int m_numToEliminate

Constant rate of attribute elimination per iteration

m_percentToEliminate

private int m_percentToEliminate

Percentage rate of attribute elimination, trumps constant rate (above threshold), ignored if = 0

m_percentThreshold

private int m_percentThreshold

Threshold below which percent elimination switches to constant elimination

m_smoCParameter

private double m_smoCParameter

Complexity parameter to pass on to SMO

m_smoTParameter

private double m_smoTParameter

Tolerance parameter to pass on to SMO

m_smoPParameter

private double m_smoPParameter

Epsilon parameter to pass on to SMO

m_smoFilterType

private int m_smoFilterType

Filter parameter to pass on to SMO

Constructor Detail