org.knime.base.node.mine.subgroupminer.apriori
Class ArrayApriori

java.lang.Object
  org.knime.base.node.mine.subgroupminer.apriori.ArrayApriori

All Implemented Interfaces:

AprioriAlgorithm

public class ArrayApriori
extends Object
implements AprioriAlgorithm

The array apriori uses the ArrayPrefixTreeNode data structure to find frequent itemsets. Based on these it constructs a prefix tree. In a prefix tree each child of an item has the path in the tree to that item in common. The path is its prefix. The transactions are processed, for each level once, by going to the node corresponding to first item in the transaction, and process the rest of the transaction for that node. Thus, there is no candidate generation.

Author:

Fabian Dill, University of Konstanz

Constructor Summary

ArrayApriori(int bitSetLength, int dbsize)
Creates an ArrayApriori instance with the bitset length, corresponding to the number of items.

Method Summary

void	findFrequentItemSets(List<BitVectorValue> transactions, double minSupport, int maxDepth, FrequentItemSet.Type type, ExecutionMonitor exec) Finds the frequent itemsets by going down the tree until the current build level is reached, there it counts those items which are present in the transaction.
List<AssociationRule>	getAssociationRules(double confidence) Returns the association rules generated from the found frequent itemsets with the passed minimal confidence.
List<FrequentItemSet>	getFrequentItemSets(FrequentItemSet.Type type) Returns the found frequent itemsets according to their type, which can either be FREE, CLOSED or MAXIMAL.
void	setMinSupport(double minSupport)

Methods inherited from class java.lang.Object

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

Constructor Detail

ArrayApriori

public ArrayApriori(int bitSetLength,
                    int dbsize)

Creates an ArrayApriori instance with the bitset length, corresponding to the number of items.

Parameters:

bitSetLength - the number of items

dbsize - the number of transactions

Method Detail

setMinSupport

public void setMinSupport(double minSupport)

Parameters:

minSupport - the minimum support

findFrequentItemSets

public void findFrequentItemSets(List<BitVectorValue> transactions,
                                 double minSupport,
                                 int maxDepth,
                                 FrequentItemSet.Type type,
                                 ExecutionMonitor exec)
                          throws CanceledExecutionException

Finds the frequent itemsets by going down the tree until the current build level is reached, there it counts those items which are present in the transaction. This implies, that it can count only those items, for which a path is present in the tree, that is, which have frequent predecessors. When the counting is finished, new children are created for those itemsets, which might become frequent in the next level, that is, itemsets with one item more. This is the method to start with when mining for frequent itemsets.

Specified by:

findFrequentItemSets in interface AprioriAlgorithm

Parameters:

transactions - a list of BitSets representing the bitvectors, thus, corresponding to the whole database

minSupport - the minimum support as an absolute value

maxDepth - the maximal length of an itemset

type - the desired type of the frequent itemsets

exec - the execution monitor

Throws:

CanceledExecutionException - if the execution was cancelled

getAssociationRules

public List<AssociationRule> getAssociationRules(double confidence)

Returns the association rules generated from the found frequent itemsets with the passed minimal confidence.

Specified by:

getAssociationRules in interface AprioriAlgorithm

Parameters:

confidence - the desired minimal confidence of the rules

Returns:

a list of association rules with the minimum confidence

getFrequentItemSets

public List<FrequentItemSet> getFrequentItemSets(FrequentItemSet.Type type)

Returns the found frequent itemsets according to their type, which can either be FREE, CLOSED or MAXIMAL.

Specified by:

getFrequentItemSets in interface AprioriAlgorithm

Parameters:

type - the desired type, either free, closed or maximal

Returns:

a list of the found frequent itemsets of the referring type