Package unifeat.featureSelection.embedded.TreeBasedMethods

Class DecisionTreeBasedMethod

java.lang.Object

unifeat.featureSelection.FeatureSelection

unifeat.featureSelection.embedded.EmbeddedApproach

unifeat.featureSelection.embedded.TreeBasedMethods.TreeBasedMethods

unifeat.featureSelection.embedded.TreeBasedMethods.DecisionTreeBasedMethod

public class DecisionTreeBasedMethod
extends TreeBasedMethods

This java class is used to implement the decision tree based methods.

Author:

Sina Tabakhi

See Also:

• TreeBasedMethods
• EmbeddedApproach
• FeatureSelection

Field Summary

Fields

Modifier and Type	Field	Description
private double	confidenceValue
private int	minNumSampleInLeaf
private int	randomTreeKValue
private int	randomTreeMaxDepth
private double	randomTreeMinNum
private double	randomTreeMinVarianceProp

Fields inherited from class unifeat.featureSelection.embedded.TreeBasedMethods.TreeBasedMethods

TEMP_PATH, TREE_TYPE

Fields inherited from class unifeat.featureSelection.embedded.EmbeddedApproach

classLabel, nameFeatures, PROJECT_PATH

Fields inherited from class unifeat.featureSelection.FeatureSelection

numClass, numFeatures, numSelectedFeature, selectedFeatureSubset, trainSet

Constructor Summary

Constructors

Constructor	Description
DecisionTreeBasedMethod(Object... arguments)	Initializes the parameters
DecisionTreeBasedMethod(String path, double confidence, int minNum)	Initializes the parameters
DecisionTreeBasedMethod(String path, int randomTreeKValue, int randomTreeMaxDepth, double randomTreeMinNum, double randomTreeMinVarianceProp)	Initializes the parameters

Method Summary

Modifier and Type	Method	Description
protected String	buildClassifier(weka.core.Instances dataTrain)	Generates a classifier using input data
void	evaluateFeatures()	Starts the feature selection process by Decision Tree based methods
protected void	selectedFeatureSubset(String tree)	Finds the feature subset from the nodes of the created tree (Used for C4.5 and Random Tree)

Methods inherited from class unifeat.featureSelection.embedded.EmbeddedApproach

loadDataSet, loadDataSet, newMethod, originalFeatureSet

Methods inherited from class unifeat.featureSelection.FeatureSelection

getSelectedFeatureSubset, setNumSelectedFeature, validate

Methods inherited from class java.lang.Object

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

Field Details

confidenceValue

private double confidenceValue

minNumSampleInLeaf

private int minNumSampleInLeaf

randomTreeKValue

private int randomTreeKValue

randomTreeMaxDepth

private int randomTreeMaxDepth

randomTreeMinNum

private double randomTreeMinNum

randomTreeMinVarianceProp

private double randomTreeMinVarianceProp

Constructor Details

DecisionTreeBasedMethod

public DecisionTreeBasedMethod(Object... arguments)

Initializes the parameters

Parameters:

arguments - array of parameter contains

if the type of tree is C4.5 (path, tree type, confidenceValue, minNumSampleInLeaf) in which path is the path of the project, tree type is the type of tree, confidenceValue is the confidence factor used for pruning, minNumSampleInLeaf is the minimum number of samples per leaf

if the type of tree is random tree (path, tree type, KValue, MaxDepth, MinNum,, MinVarianceProp) in which path is the path of the project, tree type is the type of tree, KValue is the number of randomly chosen attributes, MaxDepth is the maximum depth of the tree, MinNum is the minimum total weight of the instances in a leaf, MinVarianceProp is the minimum proportion of the total variance

DecisionTreeBasedMethod

public DecisionTreeBasedMethod(String path,
 double confidence,
 int minNum)

Initializes the parameters

Parameters:

path - the path of the project

confidence - the confidence factor used for pruning

minNum - the minimum number of samples per leaf

DecisionTreeBasedMethod

public DecisionTreeBasedMethod(String path,
 int randomTreeKValue,
 int randomTreeMaxDepth,
 double randomTreeMinNum,
 double randomTreeMinVarianceProp)

Initializes the parameters

Parameters:

path - the path of the project

randomTreeKValue - the number of randomly chosen attributes

randomTreeMaxDepth - the maximum depth of the tree

randomTreeMinNum - the minimum total weight of the instances in a leaf

randomTreeMinVarianceProp - the minimum proportion of the total variance (over all the data) required for split

Method Details

selectedFeatureSubset

protected void selectedFeatureSubset(String tree)

Finds the feature subset from the nodes of the created tree (Used for C4.5 and Random Tree)

Specified by:

selectedFeatureSubset in class TreeBasedMethods

Parameters:

tree - the generated tree based on the train set

buildClassifier

protected String buildClassifier(weka.core.Instances dataTrain)

Generates a classifier using input data

Specified by:

buildClassifier in class TreeBasedMethods

Parameters:

dataTrain - the data to train the classifier

Returns:

the output of the generated classifier

evaluateFeatures

public void evaluateFeatures()

Starts the feature selection process by Decision Tree based methods

Specified by:

evaluateFeatures in class FeatureSelection