prea.data.splitter
Class KfoldCrossValidation

java.lang.Object
  prea.data.splitter.DataSplitManager
      prea.data.splitter.KfoldCrossValidation

public class KfoldCrossValidation
extends DataSplitManager

This class implements K-fold cross-validation.

Since:

2012. 4. 20

Version:

1.1

Author:

Joonseok Lee

Field Summary

private SparseMatrix	assign
private int	foldCount

Fields inherited from class prea.data.splitter.DataSplitManager

itemCount, itemRateAverage, K_FOLD_CROSS_VALIDATION, maxValue, minValue, PREDEFINED_SPLIT, rateMatrix, SIMPLE_SPLIT, testMatrix, userCount, userRateAverage

Constructor Summary

KfoldCrossValidation(SparseMatrix originalMatrix, int k, int max, int min)
Construct an instance for K-fold cross-validation.

Method Summary

private void	divideFolds(int k) Divide the original rating matrix into k-fold.
SparseMatrix	getKthFold(int k) Return the k-th fold as test set (testMatrix), making all the others as train set in rateMatrix.

Methods inherited from class prea.data.splitter.DataSplitManager

calculateAverage, getItemRateAverage, getTestMatrix, getUserRateAverage, recoverTestItems

Methods inherited from class java.lang.Object

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

Field Detail

assign

private SparseMatrix assign

foldCount

private int foldCount

Constructor Detail

KfoldCrossValidation

public KfoldCrossValidation(SparseMatrix originalMatrix,
                            int k,
                            int max,
                            int min)

Construct an instance for K-fold cross-validation.

Method Detail

divideFolds

private void divideFolds(int k)

Divide the original rating matrix into k-fold. Illegal k would be adjusted automatically.

Parameters:

k - The index for desired fold.

getKthFold

public SparseMatrix getKthFold(int k)

Return the k-th fold as test set (testMatrix), making all the others as train set in rateMatrix.

Parameters:

k - The index for desired fold.

Returns:

Rating matrix with test data with data points in k-th fold.