org.openimaj.experiment.validation.cross

Class KFold<INSTANCE>

java.lang.Object

org.openimaj.experiment.validation.cross.KFold<INSTANCE>

Type Parameters:

INSTANCE - Type of instances

All Implemented Interfaces:

CrossValidator<ListDataset<INSTANCE>>

public class KFold<INSTANCE>
extends Object
implements CrossValidator<ListDataset<INSTANCE>>

K-Fold cross validation for ListDatasets. The data is broken into K approximately equally sized non-overlapping randomised subsets. On each iteration, one subset is picked as the validation data and the remaining subsets are combined to make the training data. The number of iterations is equal to the number of subsets.

If the number of subsets is equal to the number of instances, then the K-Fold Cross Validation scheme becomes equivalent to the LOOCV scheme. The implementation of LOOCV in the LeaveOneOut class is considerably more memory efficient than using this class however.

Author:

Jonathon Hare (jsh2@ecs.soton.ac.uk)

Constructor Summary

Constructors

Constructor and Description
KFold(int k) Construct with the given number of folds.

Method Summary

Modifier and Type	Method and Description
CrossValidationIterable<ListDataset<INSTANCE>>	createIterable(ListDataset<INSTANCE> data) Create the iterable from the dataset.
String	toString()

Methods inherited from class java.lang.Object

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

Constructor Detail

KFold

public KFold(int k)

Construct with the given number of folds.

Parameters:

k - the number of folds.

Method Detail

createIterable

public CrossValidationIterable<ListDataset<INSTANCE>> createIterable(ListDataset<INSTANCE> data)

Description copied from interface: CrossValidator

Create the iterable from the dataset.

Specified by:

createIterable in interface CrossValidator<ListDataset<INSTANCE>>

Parameters:

data - the dataset

Returns:

the iterable

toString

public String toString()

Overrides:

toString in class Object