ca.uottawa.balie
Class WekaAttributeSelection

java.lang.Object
  ca.uottawa.balie.WekaAttributeSelection

public class WekaAttributeSelection
extends java.lang.Object

Methods to select the top attributes from a given classification problem.

Author:

nadeaud

Field Summary

static int	DEFAULT_NUM_ATTRIBUTES
static int	WEKA_CHI_SQUARE
static int	WEKA_INFO_GAIN
static int	WEKA_ONER
static int	WEKA_RELIEF
static int	WEKA_SVM
static int	WEKA_SYM_UNCERT
static int	WEKA_WRAPPER

Constructor Summary

WekaAttributeSelection(int pi_Evaluator, WekaAttribute[] pi_Attributes, java.lang.String[] pi_ClassAttributes)
Creates an Attribute selection framework.

Method Summary

void	AddInstance(java.lang.Object[] pi_Instance, java.lang.String pi_Class) Add an instance, that is a vector of value for all attributes.
weka.core.FastVector	GetAttribute()
java.lang.String[]	GetAttributeList()
static void	main(java.lang.String[] args) Test routine.
void	NumAttributes(int pi_Num) Sets the number of top attribute to select.
java.lang.String[]	ReduceDimentionality() Gets the array of chosen attributes
void	Select(boolean pi_Debug) Select the top attributes
void	SetDoubleOnly(boolean b)

Methods inherited from class java.lang.Object

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

Field Detail

WEKA_CHI_SQUARE

public static final int WEKA_CHI_SQUARE

See Also:

Constant Field Values

WEKA_INFO_GAIN

public static final int WEKA_INFO_GAIN

See Also:

Constant Field Values

WEKA_WRAPPER

public static final int WEKA_WRAPPER

See Also:

Constant Field Values

WEKA_SYM_UNCERT

public static final int WEKA_SYM_UNCERT

See Also:

Constant Field Values

WEKA_SVM

public static final int WEKA_SVM

See Also:

Constant Field Values

WEKA_RELIEF

public static final int WEKA_RELIEF

See Also:

Constant Field Values

WEKA_ONER

public static final int WEKA_ONER

See Also:

Constant Field Values

DEFAULT_NUM_ATTRIBUTES

public static final int DEFAULT_NUM_ATTRIBUTES

See Also:

Constant Field Values

Constructor Detail

WekaAttributeSelection

public WekaAttributeSelection(int pi_Evaluator,
                              WekaAttribute[] pi_Attributes,
                              java.lang.String[] pi_ClassAttributes)

Creates an Attribute selection framework.

Parameters:

pi_Evaluator - Type of evaluation of attributes merit

pi_Attributes - The attributes to evaluate

pi_ClassAttributes - The class attribute

Method Detail

AddInstance

public void AddInstance(java.lang.Object[] pi_Instance,
                        java.lang.String pi_Class)

Add an instance, that is a vector of value for all attributes.

Parameters:

pi_Instance - The instance, an array of objects (can mix numeric and nominal attributes - see WekaAttribute)

pi_Class - The class for this instance

See Also:

WekaAttribute

NumAttributes

public void NumAttributes(int pi_Num)

Sets the number of top attribute to select.

Parameters:

pi_Num - Num of attributes to select

GetAttributeList

public java.lang.String[] GetAttributeList()

GetAttribute

public weka.core.FastVector GetAttribute()

SetDoubleOnly

public void SetDoubleOnly(boolean b)

Select

public void Select(boolean pi_Debug)

Select the top attributes

ReduceDimentionality

public java.lang.String[] ReduceDimentionality()

Gets the array of chosen attributes

Returns:

A String array where each element is the label of a chosen attribute (see WekaAttribute).

See Also:

WekaAttribute

main

public static void main(java.lang.String[] args)

Test routine.

Parameters:

args -