keel.Algorithms.UnsupervisedLearning.AssociationRules.IntervalRuleLearning.FPgrowth.LUCS_KDD

Class TotalSupportTree

java.lang.Object

keel.Algorithms.UnsupervisedLearning.AssociationRules.IntervalRuleLearning.FPgrowth.LUCS_KDD.AssocRuleMining

keel.Algorithms.UnsupervisedLearning.AssociationRules.IntervalRuleLearning.FPgrowth.LUCS_KDD.TotalSupportTree

Direct Known Subclasses:

FPtree

public class TotalSupportTree
extends AssocRuleMining

Since:

JDK1.5

Version:

1.1

Author:

Modified by Nicolò Flugy Papè (Politecnico di Milano) 24/03/2009

Nested Class Summary

Nested classes/interfaces inherited from class keel.Algorithms.UnsupervisedLearning.AssociationRules.IntervalRuleLearning.FPgrowth.LUCS_KDD.AssocRuleMining

AssocRuleMining.RuleNode

Field Summary

Fields

Modifier and Type	Field and Description
protected long	numUpdates The number of updates required to generate the T-tree.
protected TtreeNode[]	startTtreeRef The reference to start of t-tree.

Fields inherited from class keel.Algorithms.UnsupervisedLearning.AssociationRules.IntervalRuleLearning.FPgrowth.LUCS_KDD.AssocRuleMining

confidence, conversionArray, dataArray, dataset, isOrderedFlag, isPrunedFlag, minSupport, numCols, numOneItemSets, numRows, reconversionArray, startRulelist, support

Constructor Summary

Constructors

Constructor and Description
TotalSupportTree(myDataset ds, double sup, double conf) Constructor to process dataset and parameters.

Method Summary

Modifier and Type	Method and Description
void	addToTtree(short[] itemSet, int support) Commences process of adding an itemset (with its support value) to a T-tree when using a T-tree either as a storage mechanism, or when adding to an existing T-tree.
protected TtreeNode[]	addToTtree(TtreeNode[] linkRef, int size, int endIndex, short[] itemSet, int support) Inserts a node into a T-tree.
protected int	countNumFreqSets() Commences process of counting the number of frequent (large/supported sets contained in the T-tree.
protected int	countNumFreqSets(int size, TtreeNode[] linkRef, int num) Counts the number of supported nodes in a sub branch of the T-tree.
void	generateARs() Initiates process of generating Association Rules (ARs) from a T-tree.
protected void	generateARs(short[] itemSetSofar, int size, TtreeNode[] linkRef) Continues process of generating association rules from a T-tree by recursively looping through T-tree level by level.
protected void	generateARs2() Loops through top level of T-tree as part of the AR generation process.
java.util.HashSet<java.lang.Integer>	getCoveredRecords() Retrieves all the records which are covered by the association rules
protected java.util.HashSet<java.lang.Integer>	getCoveredRecords(short[] itemset)
int	getNumFreqSets() Commences the process of counting and returning number of supported nodes in the T-tree.
protected int	getSupportForItemSetInTtree(short[] itemSet) Commences process for finding the support value for the given item set in the T-tree (which is know to exist in the T-tree).
void	outputFrequentSets() Commences the process of outputting the frequent sets contained in the T-tree.
void	setNumOneItemSets() Sets the number of one item sets field (numOneItemSets to the number of supported one item sets.

Methods inherited from class keel.Algorithms.UnsupervisedLearning.AssociationRules.IntervalRuleLearning.FPgrowth.LUCS_KDD.AssocRuleMining

complement, copyItemSet, countSingles, defConvertArrays, getNumSupOneItemSets, getRulesSet, idInputDataOrdering, insertRuleintoRulelist, notMemberOf, orderFirstNofCountArray, outputDataArray, outputItemSet, outputRules, outputRules, realloc1, realloc2, reallocInsert, recastInputData, recastInputDataAndPruneUnsupportedAtts, reconvertItem, reconvertItemSet, removeElementN, sortItemSet, twoDecPlaces

Methods inherited from class java.lang.Object

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

Field Detail

startTtreeRef

protected TtreeNode[] startTtreeRef

The reference to start of t-tree.

numUpdates

protected long numUpdates

The number of updates required to generate the T-tree.

Constructor Detail

TotalSupportTree

public TotalSupportTree(myDataset ds,
                        double sup,
                        double conf)

Constructor to process dataset and parameters.

Parameters:

ds - The instance of the dataset for dealing with its records

sup - The user-specified minimum support for the mined association rules

conf - The user-specified minimum confidence for the mined association rules

Method Detail

addToTtree

public void addToTtree(short[] itemSet,
                       int support)

Commences process of adding an itemset (with its support value) to a T-tree when using a T-tree either as a storage mechanism, or when adding to an existing T-tree.

Parameters:

itemSet - The given itemset. Listed in numeric order (not reverse numeric order!).

support - The support value associated with the given itemset.

addToTtree

protected TtreeNode[] addToTtree(TtreeNode[] linkRef,
                                 int size,
                                 int endIndex,
                                 short[] itemSet,
                                 int support)

Inserts a node into a T-tree.

Recursive procedure.

Parameters:

linkRef - the reference to the current array in Ttree.

size - the size of the current array in T-tree.

endIndex - the index of the last element/attribute in the itemset, which is also used as a level counter.

itemSet - the given itemset.

support - the support value associated with the given itemset.

Returns:

the reference to the revised sub-branch of t-tree.

getSupportForItemSetInTtree

protected int getSupportForItemSetInTtree(short[] itemSet)

Commences process for finding the support value for the given item set in the T-tree (which is know to exist in the T-tree).

Used when generating Association Rules (ARs). Note that itemsets are stored in reverse order in the T-tree therefore the given itemset must be processed in reverse.

Parameters:

itemSet - the given itemset.

Returns:

returns the support value (0 if not found).

generateARs

public void generateARs()

Initiates process of generating Association Rules (ARs) from a T-tree.

generateARs2

protected void generateARs2()

Loops through top level of T-tree as part of the AR generation process.

generateARs

protected void generateARs(short[] itemSetSofar,
                           int size,
                           TtreeNode[] linkRef)

Continues process of generating association rules from a T-tree by recursively looping through T-tree level by level.

Parameters:

itemSetSofar - the label for a T-tree node as generated sofar.

size - the length/size of the current array lavel in the T-tree.

linkRef - the reference to the current array level in the T-tree.

setNumOneItemSets

public void setNumOneItemSets()

Sets the number of one item sets field (numOneItemSets to the number of supported one item sets.

outputFrequentSets

public void outputFrequentSets()

Commences the process of outputting the frequent sets contained in the T-tree.

countNumFreqSets

protected int countNumFreqSets()

Commences process of counting the number of frequent (large/supported sets contained in the T-tree.

Returns:

number of frequent sets counted

countNumFreqSets

protected int countNumFreqSets(int size,
                               TtreeNode[] linkRef,
                               int num)

Counts the number of supported nodes in a sub branch of the T-tree.

Parameters:

size - the length/size of the current array level in the T-tree.

linkRef - the reference to the current array level in the T-tree.

num - the number of frequent sets sofar.

Returns:

the counted number of supported nodes in a sub branch

getNumFreqSets

public int getNumFreqSets()

Commences the process of counting and returning number of supported nodes in the T-tree.

A supported set is assumed to be a non null node in the T-tree.

Returns:

the number of supported nodes in the T-tree.

getCoveredRecords

protected java.util.HashSet<java.lang.Integer> getCoveredRecords(short[] itemset)

getCoveredRecords

public java.util.HashSet<java.lang.Integer> getCoveredRecords()

Retrieves all the records which are covered by the association rules

Returns:

a set of TIDs representing the covered records.