it.uniroma2.sag.kelp.data.manipulator

Class TreePairRelTagger

java.lang.Object

it.uniroma2.sag.kelp.data.manipulator.TreePairRelTagger

All Implemented Interfaces:

Manipulator

public class TreePairRelTagger
extends Object
implements Manipulator

This manipulator establishes relations between two tree representations. Given two trees whose nodes are \(N_A=\left \{ n_1^A, n_2^A... \right \}) and \(N_B=\left \{ n_1^B, n_2^B... \right \}), the manipulators firstly selects two subsets \(S_A \subseteq N_A\) and \(S_B \subseteq N_B\) according to the filtering policy defined by a TreeNodeFilter. Then the nodes \(n_i^A \in S_A\) and the nodes \(n_i^B \in S_B\) become anchors if their similarity evaluated with nodeSimilarity is higher than simThreshold. Finally anchor nodes are marked using a markingPolicy and the same marking is applied to their ascendants (till upwardPropagation generations) and descendants (till downPropagation generations).

Author:

Simone Filice

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	TreePairRelTagger.MARKING_POLICY The marking policy applied by TreePairRelTagger

Field Summary

Fields

Modifier and Type	Field and Description
static String	DEFAULT_PREFIX

Constructor Summary

Constructors

Constructor and Description
TreePairRelTagger(int upwardPropagation, int downwordPropagation, String representation, TreeNodeFilter nodeFilter, TreePairRelTagger.MARKING_POLICY markingPolicy, TreeNodeSimilarity nodeSimilarity, float simThreshold) Constructor for TreePairRelTagger

Method Summary

Methods

Modifier and Type	Method and Description
void	establishNodeRelations(TreeRepresentation treeA, TreeRepresentation treeB) Establishes relations between treeA and treeA Let \(N_A=\left \{ n_1^A, n_2^A...
String	getMarkingPrefix() Returns the prefix used to mark the related nodes
void	manipulate(Example example)
void	setMarkingPrefix(String markingPrefix) Sets the prefix used to mark the related nodes

Methods inherited from class java.lang.Object

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

Field Detail

DEFAULT_PREFIX

public static final String DEFAULT_PREFIX

See Also:

Constant Field Values

Constructor Detail

TreePairRelTagger

public TreePairRelTagger(int upwardPropagation,
                 int downwordPropagation,
                 String representation,
                 TreeNodeFilter nodeFilter,
                 TreePairRelTagger.MARKING_POLICY markingPolicy,
                 TreeNodeSimilarity nodeSimilarity,
                 float simThreshold)

Constructor for TreePairRelTagger

NOTE: the related nodes will be marked using DEFAULT_PREFIX. To specify an alternative prefix the setMarkingPrefix can be used

Parameters:

upwardPropagation - the number of generations of ascendants to which the mark is propagated

downwordPropagation - the number of generations of descendants to which the mark is propagated

representation - the name of the tree representations that must be related

nodeFilter - the policy to be applied in selecting potential anchor nodes

markingPolicy - the marking policy specifying what kind of mark must be applied

nodeSimilarity - the similarity metric between tree nodes

simThreshold - the similarity threshold that two nodes should overcome to be considered anchors (and then marked)

Method Detail

getMarkingPrefix

public String getMarkingPrefix()

Returns the prefix used to mark the related nodes

Returns:

the markingPrefix

setMarkingPrefix

public void setMarkingPrefix(String markingPrefix)

Sets the prefix used to mark the related nodes

Parameters:

markingPrefix - the markingPrefix to set

manipulate

public void manipulate(Example example)

Specified by:

manipulate in interface Manipulator

establishNodeRelations

public void establishNodeRelations(TreeRepresentation treeA,
                          TreeRepresentation treeB)

Establishes relations between treeA and treeA Let \(N_A=\left \{ n_1^A, n_2^A... \right \}) be the nodes of treeA and \(N_B=\left \{ n_1^B, n_2^B... \right \}) be the nodes of treeB, the method firstly selects two subsets \(S_A \subseteq N_A\) and \(S_B \subseteq N_B\) according to the filtering policy defined by a TreeNodeFilter. Then the nodes \(n_i^A \in S_A\) and the nodes \(n_i^B \in S_B\) become anchors if their similarity evaluated with nodeSimilarity is higher than simThreshold. Finally anchor nodes are marked using a markingPolicy and the same marking is applied to their ascendants (till upwardPropagation generations) and descendants (till downPropagation generations).

Parameters:

treeA -

treeB -