ai.libs.jaicore.ml.cache

Class ReproducibleInstances

java.lang.Object

java.util.AbstractCollection<E>

java.util.AbstractList<weka.core.Instance>

weka.core.Instances

ai.libs.jaicore.ml.cache.ReproducibleInstances

All Implemented Interfaces:

java.io.Serializable, java.lang.Iterable<weka.core.Instance>, java.util.Collection<weka.core.Instance>, java.util.List<weka.core.Instance>, weka.core.RevisionHandler

public class ReproducibleInstances
extends weka.core.Instances

New Instances class to track splits and data origin. Origin of the dataset is stored by a LoadDataSetInstruction and changed by FoldBasedSubsetInstructions saved as a list of instructions. This history of the instances can be converted to json and used to reproduce a specific set of instances.

See Also:

Serialized Form

Field Summary

Fields inherited from class weka.core.Instances

ARFF_DATA, ARFF_RELATION, FILE_EXTENSION, m_Attributes, m_ClassIndex, m_Instances, m_Lines, m_NamesToAttributeIndices, m_RelationName, SERIALIZED_OBJ_FILE_EXTENSION

Fields inherited from class java.util.AbstractList

modCount

Constructor Summary

Constructors

Constructor and Description
ReproducibleInstances(ReproducibleInstances dataset)

Method Summary

Modifier and Type	Method and Description
void	addInstruction(Instruction i) Adds a new Instruction to the history of these Instances
static ReproducibleInstances	fromARFF(java.lang.String path) Creates a new ReproducibleInstances object.
static ReproducibleInstances	fromHistory(java.util.List<Instruction> history, java.lang.String apiKey) Creates a ReproducibleInstances Object based on the given History.
static ReproducibleInstances	fromOpenML(java.lang.String id, java.lang.String apiKey) Creates a new ReproducibleInstances object.
java.util.List<Instruction>	getInstructions()
boolean	isCacheLookup() If true signifies that performance on this data should be looked up in cache
boolean	isCacheStorage() If true signifies that performance evaluation should be stored.
void	setCacheLookup(boolean cacheLookup) If true signifies that performance on this data should be looked up in cache
void	setCacheStorage(boolean cacheStorage) If set to true, signifies that performance evaluation should be stored.

Methods inherited from class weka.core.Instances

add, add, allAttributeWeightsIdentical, allInstanceWeightsIdentical, attribute, attribute, attributeStats, attributeToDoubleArray, checkForAttributeType, checkForStringAttributes, checkInstance, classAttribute, classIndex, compactify, copyInstances, delete, delete, deleteAttributeAt, deleteAttributeType, deleteStringAttributes, deleteWithMissing, deleteWithMissing, deleteWithMissingClass, enumerateAttributes, enumerateInstances, equalHeaders, equalHeadersMsg, firstInstance, get, getRandomNumberGenerator, getRevision, initialize, insertAttributeAt, instance, instancesAndWeights, kthSmallestValue, kthSmallestValue, lastInstance, main, meanOrMode, meanOrMode, mergeInstances, numAttributes, numClasses, numDistinctValues, numDistinctValues, numInstances, randomize, readInstance, relationName, remove, renameAttribute, renameAttribute, renameAttributeValue, renameAttributeValue, replaceAttributeAt, resample, resampleWithWeights, resampleWithWeights, resampleWithWeights, resampleWithWeights, resampleWithWeights, resampleWithWeights, resampleWithWeights, set, setAttributeWeight, setAttributeWeight, setClass, setClassIndex, setRelationName, size, sort, sort, sortBasedOnNominalAttribute, stableSort, stableSort, stratify, stratStep, stringFreeStructure, stringWithoutHeader, sumOfWeights, swap, test, testCV, toString, toSummaryString, trainCV, trainCV, variance, variance, variances

Methods inherited from class java.util.AbstractList

addAll, clear, equals, hashCode, indexOf, iterator, lastIndexOf, listIterator, listIterator, removeRange, subList

Methods inherited from class java.util.AbstractCollection

addAll, contains, containsAll, isEmpty, remove, removeAll, retainAll, toArray, toArray

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface java.util.List

addAll, contains, containsAll, isEmpty, remove, removeAll, replaceAll, retainAll, sort, spliterator, toArray, toArray

Methods inherited from interface java.util.Collection

parallelStream, removeIf, stream

Methods inherited from interface java.lang.Iterable

forEach

Constructor Detail

ReproducibleInstances

public ReproducibleInstances(ReproducibleInstances dataset)

Method Detail

fromHistory

public static ReproducibleInstances fromHistory(java.util.List<Instruction> history,
                                                java.lang.String apiKey)
                                         throws java.io.IOException

Creates a ReproducibleInstances Object based on the given History. Instructions that no not modify the Instances will be ignored (No evaluation will be done).

Parameters:

history - List of Instructions used to create the original Instances

apiKey - apiKey in case openml.org is used

Returns:

new ReproducibleInstances object

Throws:

java.io.IOException - if something goes wrong while loading Instances from openml or when reading arff file

fromOpenML

public static ReproducibleInstances fromOpenML(java.lang.String id,
                                               java.lang.String apiKey)
                                        throws java.io.IOException

Creates a new ReproducibleInstances object. Data is loaded from openml.org.

Parameters:

id - The id of the openml dataset

apiKey - apikey to use

Returns:

new ReproducibleInstances object

Throws:

java.io.IOException - if something goes wrong while loading Instances from openml

fromARFF

public static ReproducibleInstances fromARFF(java.lang.String path)
                                      throws java.io.IOException

Creates a new ReproducibleInstances object. Data is loaded from a local arff file.

Parameters:

path - path to the dataset

Returns:

new ReproducibleInstances object

Throws:

java.io.IOException - if the ARFF file is not read successfully

getInstructions

public java.util.List<Instruction> getInstructions()

Returns:

the ordered lists of instructions or null if cache is not used

addInstruction

public void addInstruction(Instruction i)

Adds a new Instruction to the history of these Instances

Parameters:

i - - new Instruction

isCacheStorage

public boolean isCacheStorage()

If true signifies that performance evaluation should be stored.

Returns:

true if performance should be saved

setCacheStorage

public void setCacheStorage(boolean cacheStorage)

If set to true, signifies that performance evaluation should be stored.

Parameters:

cacheStorage - the cacheStorage to set

isCacheLookup

public boolean isCacheLookup()

If true signifies that performance on this data should be looked up in cache

Returns:

true if lookup should be performed

setCacheLookup

public void setCacheLookup(boolean cacheLookup)

If true signifies that performance on this data should be looked up in cache

Parameters:

cacheLookup - the cacheLookup to set