ai.libs.jaicore.ml.weka.dataset.splitter

Class MultilabelDatasetSplitter

java.lang.Object

ai.libs.jaicore.ml.weka.dataset.splitter.MultilabelDatasetSplitter

public class MultilabelDatasetSplitter
extends java.lang.Object

This class provides methods to obtain train and test splits for a given data set and split technique.

Method Summary

Modifier and Type	Method and Description
static java.lang.String	getSplitTechniqueAndDetailsSeparator() Obtain the token used to separate a split technique and the details about the split.
static weka.core.Instances	getTestSplit(weka.core.Instances data, java.lang.String splitDescription, java.lang.String testFold, java.lang.String seed) Split the Instances object according to the given splitDescription.
static weka.core.Instances	getTrainSplit(weka.core.Instances data, java.lang.String splitDescription, java.lang.String testFold, java.lang.String seed) Split the Instances object according to the given splitDescription.

Methods inherited from class java.lang.Object

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

Method Detail

getTestSplit

public static weka.core.Instances getTestSplit(weka.core.Instances data,
                                               java.lang.String splitDescription,
                                               java.lang.String testFold,
                                               java.lang.String seed)

Split the Instances object according to the given splitDescription. The splitDescription is composed of the used technique and details about the split for the used technique separated by special token obtainable by getSplitTechniqueAndDetailsSeparator(). The returned data will only contain the testFold. The seed is given to the technique if it uses a seed.

Parameters:

data - The data to extract a test fold from

splitDescription - The description of how the split shall be performed

testFold - The number of the fold to be extracted (0...n)

seed - The random seed to be used by the technique

Returns:

The test fold

getTrainSplit

public static weka.core.Instances getTrainSplit(weka.core.Instances data,
                                                java.lang.String splitDescription,
                                                java.lang.String testFold,
                                                java.lang.String seed)

Split the Instances object according to the given splitDescription. The splitDescription is composed of the used technique and details about the split for the used technique separated by a special token obtainable by getSplitTechniqueAndDetailsSeparator(). The returned data will not contain the testFold. The seed is given to the technique if it uses a seed.

Parameters:

data - The data to extract a train fold from

splitDescription - The description of how the split shall be performed

testFold - The number of the fold to be excluded (0..n)

seed - The random seed to be used by the technique

Returns:

The train fold

getSplitTechniqueAndDetailsSeparator

public static java.lang.String getSplitTechniqueAndDetailsSeparator()

Obtain the token used to separate a split technique and the details about the split.

Returns:

The separator token