Package unifeat.classifier.evaluation.wekaClassifier

Class CrossValidation

java.lang.Object

unifeat.classifier.evaluation.wekaClassifier.CrossValidation

public class CrossValidation
extends Object

This java class is used to apply the classifiers for computing the performance of the feature selection methods. The classifiers have been implemented as the Weka software. The k-fold cross validation method is used for evaluating classifiers.

Author:

Sina Tabakhi

Constructor Summary

Constructors

Constructor	Description
CrossValidation()

Method Summary

Modifier and Type	Method	Description
static Criteria	dTree(String pathTrainData, double confidenceValue, int minNumSampleInLeaf, int kFold)	This method builds and evaluates the decision tree(DT) classifier.
static Criteria	kNN(String pathTrainData, int kNNValue, int kFold)	This method builds and evaluates the k-nearest neighbours(knn) classifier.
static Criteria	naiveBayes(String pathTrainData, int kFold)	This method builds and evaluates the naiveBayes(NB) classifier.
static Criteria	SVM(String pathTrainData, SVMKernelType svmKernel, double c, int kFold)	This method builds and evaluates the support vector machine(SVM) classifier.

Methods inherited from class java.lang.Object

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

Constructor Details

CrossValidation

public CrossValidation()

Method Details

SVM

public static Criteria SVM(String pathTrainData,
 SVMKernelType svmKernel,
 double c,
 int kFold)

This method builds and evaluates the support vector machine(SVM) classifier. The SMO is used as the SVM classifier implemented in the Weka software.

Parameters:

pathTrainData - the path of the train set

svmKernel - the kernel to use

c - the complexity parameter C

kFold - the number of equal sized subsamples

Returns:

the different criteria values

See Also:

• Criteria

naiveBayes

public static Criteria naiveBayes(String pathTrainData,
 int kFold)

This method builds and evaluates the naiveBayes(NB) classifier. The naiveBayes is used as the NB classifier implemented in the Weka software.

Parameters:

pathTrainData - the path of the train set

kFold - the number of equal sized subsamples

Returns:

the different criteria values

See Also:

• Criteria

dTree

public static Criteria dTree(String pathTrainData,
 double confidenceValue,
 int minNumSampleInLeaf,
 int kFold)

This method builds and evaluates the decision tree(DT) classifier. The j48 is used as the DT classifier implemented in the Weka software.

Parameters:

pathTrainData - the path of the train set

confidenceValue - The confidence factor used for pruning

minNumSampleInLeaf - The minimum number of instances per leaf

kFold - the number of equal sized subsamples

Returns:

the different criteria values

See Also:

• Criteria

kNN

public static Criteria kNN(String pathTrainData,
 int kNNValue,
 int kFold)

This method builds and evaluates the k-nearest neighbours(knn) classifier. The IBk is used as the knn classifier implemented in the Weka software.

Parameters:

pathTrainData - the path of the train set

kNNValue - the number of neighbours to use

kFold - the number of equal sized subsamples

Returns:

the different criteria values

See Also:

• Criteria