edu.udo.cs.yale.example
Class SplittedExampleSet

java.lang.Object
  edu.udo.cs.yale.operator.AbstractIOObject
      edu.udo.cs.yale.operator.ResultObjectAdapter
          edu.udo.cs.yale.example.AbstractExampleSet
              edu.udo.cs.yale.example.ExampleSetAdapter
                  edu.udo.cs.yale.example.SplittedExampleSet

All Implemented Interfaces:

ExampleSet, IOObject, ResultObject, java.lang.Cloneable, java.lang.Iterable<Example>

public class SplittedExampleSet
extends ExampleSetAdapter

An example set that can be split into subsets by using a Partition.

Version:

$Id: SplittedExampleSet.java,v 2.26 2006/03/23 17:48:24 ingomierswa Exp $

Author:

Simon Fischer, Ingo Mierswa

Field Summary

static int	LINEAR_SAMPLING Indicates a non-shuffled sampling for partition building.
private Partition	partition The partition.
static java.lang.String[]	SAMPLING_NAMES Indicates a non-shuffled sampling for partition building.
static int	SHUFFLED_SAMPLING Indicates a shuffled sampling for partition building.
static int	STRATIFIED_SAMPLING Indicates a stratified shuffled sampling for partition building.

Fields inherited from class edu.udo.cs.yale.example.ExampleSetAdapter

delegate

Fields inherited from interface edu.udo.cs.yale.example.ExampleSet

ATTRIBUTE_NAME, BATCH_NAME, CLUSTER_NAME, CONFIDENCE_NAME, ID_NAME, KNOWN_ATTRIBUTE_TYPES, LABEL_NAME, PREDICTION_NAME, TYPE_ATTRIBUTE, TYPE_BATCH, TYPE_CLUSTER, TYPE_ID, TYPE_LABEL, TYPE_PREDICTION, TYPE_WEIGHT, WEIGHT_NAME

Constructor Summary

SplittedExampleSet(ExampleSet exampleSet, double[] splitRatios, int samplingType, int seed)
Creates an example set that is splitted into n subsets with the given sampling type.

SplittedExampleSet(ExampleSet exampleSet, double splitRatio, int samplingType, int seed)
Creates an example set that is splitted into two subsets using the given sampling type.

SplittedExampleSet(ExampleSet exampleSet, int numberOfSubsets, int samplingType, int seed)
Creates an example set that is splitted into numberOfSubsets parts with the given sampling type.

SplittedExampleSet(ExampleSet exampleSet, Partition partition)
Constructs a SplittedExampleSet with the given partition.

SplittedExampleSet(SplittedExampleSet exampleSet)
Clone constructor.

Method Summary

void	clearSelection() Clears the current selection
private static PartitionBuilder	createPartitionBuilder(ExampleSet exampleSet, int samplingType, int seed) Creates the partition builder for the given sampling type.
int	getActualParentIndex(int index) Returns the index of the example in the parent example set.
Example	getExample(int index) Searches i-th example in the currently selected partition.
int	getNumberOfSubsets() Returns the number of subsets.
void	invertSelection() Inverts the current selection
java.util.Iterator<Example>	iterator() Returns an example reader that splits all examples that are not selected.
void	selectAllSubsetsBut(int index) Selects all but one subset.
void	selectSingleSubset(int index) Selects exactly one subset.
int	size() Returns the number of examples in this example set.
static SplittedExampleSet	splitByAttribute(ExampleSet exampleSet, Attribute attribute) Works only for nominal and integer attributes.

Methods inherited from class edu.udo.cs.yale.example.ExampleSetAdapter

addAttribute, clearUserData, getAttribute, getAttribute, getExampleFromId, getExampleReader, getExampleTable, getNumberOfAttributes, getSize, getSpecialAttribute, getSpecialAttributeNames, getSpecialAttributes, getUserData, getUserDataKeys, remapAllIds, removeAttribute, replaceAttribute, setSpecialAttribute, setUserData

Methods inherited from class edu.udo.cs.yale.example.AbstractExampleSet

addAllAttributes, clearPredictedLabel, clearWeight, clone, contains, containsValueType, copy, createClusterAttribute, createDataTable, createSpecialAttribute, createWeightAttribute, equals, getActions, getBlockEndIndex, getCluster, getId, getLabel, getName, getPredictedLabel, getVisualisationComponent, getWeight, recalculateAllAttributeStatistics, recalculateAttributeStatistics, recalculateAttributeStatistics, removeAllAttributes, removeAttribute, setAttributes, setCluster, setId, setLabel, setPredictedLabel, setWeight, toString, writeAttributeFile, writeDataFile, writeSparseAttributeFile, writeSparseDataFile

Methods inherited from class edu.udo.cs.yale.operator.ResultObjectAdapter

addAction, toHTML, toResultString

Methods inherited from class edu.udo.cs.yale.operator.AbstractIOObject

read, read, read, write, write

Methods inherited from class java.lang.Object

finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Methods inherited from interface edu.udo.cs.yale.operator.ResultObject

toResultString

Methods inherited from interface edu.udo.cs.yale.operator.IOObject

write, write

Field Detail

SAMPLING_NAMES

public static final java.lang.String[] SAMPLING_NAMES

Indicates a non-shuffled sampling for partition building.

LINEAR_SAMPLING

public static final int LINEAR_SAMPLING

Indicates a non-shuffled sampling for partition building.

See Also:

Constant Field Values

SHUFFLED_SAMPLING

public static final int SHUFFLED_SAMPLING

Indicates a shuffled sampling for partition building.

See Also:

Constant Field Values

STRATIFIED_SAMPLING

public static final int STRATIFIED_SAMPLING

Indicates a stratified shuffled sampling for partition building.

See Also:

Constant Field Values

partition

private Partition partition

The partition.

Constructor Detail

SplittedExampleSet

public SplittedExampleSet(SplittedExampleSet exampleSet)

Clone constructor.

SplittedExampleSet

public SplittedExampleSet(ExampleSet exampleSet,
                          Partition partition)

Constructs a SplittedExampleSet with the given partition.

SplittedExampleSet

public SplittedExampleSet(ExampleSet exampleSet,
                          double splitRatio,
                          int samplingType,
                          int seed)

Creates an example set that is splitted into two subsets using the given sampling type.

SplittedExampleSet

public SplittedExampleSet(ExampleSet exampleSet,
                          double[] splitRatios,
                          int samplingType,
                          int seed)

Creates an example set that is splitted into n subsets with the given sampling type.

SplittedExampleSet

public SplittedExampleSet(ExampleSet exampleSet,
                          int numberOfSubsets,
                          int samplingType,
                          int seed)

Creates an example set that is splitted into numberOfSubsets parts with the given sampling type.

Method Detail

createPartitionBuilder

private static PartitionBuilder createPartitionBuilder(ExampleSet exampleSet,
                                                       int samplingType,
                                                       int seed)

Creates the partition builder for the given sampling type.

selectSingleSubset

public void selectSingleSubset(int index)

Selects exactly one subset.

selectAllSubsetsBut

public void selectAllSubsetsBut(int index)

Selects all but one subset.

invertSelection

public void invertSelection()

Inverts the current selection

clearSelection

public void clearSelection()

Clears the current selection

getNumberOfSubsets

public int getNumberOfSubsets()

Returns the number of subsets.

iterator

public java.util.Iterator<Example> iterator()

Returns an example reader that splits all examples that are not selected.

Specified by:

iterator in interface java.lang.Iterable<Example>

Overrides:

iterator in class ExampleSetAdapter

size

public int size()

Description copied from interface: ExampleSet

Returns the number of examples in this example set. This number should not be used to create for-loops to iterate through all examples.

Specified by:

size in interface ExampleSet

Overrides:

size in class ExampleSetAdapter

getExample

public Example getExample(int index)

Searches i-th example in the currently selected partition. This is done in constant time.

Specified by:

getExample in interface ExampleSet

Overrides:

getExample in class ExampleSetAdapter

getActualParentIndex

public int getActualParentIndex(int index)

Returns the index of the example in the parent example set.

splitByAttribute

public static SplittedExampleSet splitByAttribute(ExampleSet exampleSet,
                                                  Attribute attribute)

Works only for nominal and integer attributes. If k is the number of different values, this method splits the example set into k subsets according to the value of the given attribute.