Interface: HotspotIdentifier

Interface class: de.dfki.km.phaselib.impl.alignments.hotspot.HotspotIdentifier

The purpose of a HotspotIdentifier is the identification of hotspots ;-) More specifically, it is to find two classes that are a "sure match" , i.e. that are very likely to get aligned by an equality relation, and then wrap them into a simple Hotspot object.

The HotspotIdentifier should keep track of the hotspots already returned, since the main algorithm will not recognise redundantly returned hotspots and do unnecessary work. Furthermore, a sophisticated implementation should take into account the current situation (hotspots, sizes, coverage) when looking for new hotspots. For example, they tend to be in the vicinity of other (high rated) ones.

Implementations

HotspotIdentifier.PURELY_RANDOM

This implementation selects the two anchor classes purely by chance. This may be, however, a reasonable approach when the equality of two singled out entities is not easily detectable.

Parameters

Parameter name	ValueType	Default	Description
PARAM_USE_GENERATOR	AlignmentGenerator	SimpleEvidenceMatcher(StringBasedSimilarity)	Defines the AlignmentGenerator to be used in the Hotspots.

NameSimilarityIdentifier

This HotspotIdentifier uses simple string matching to find equally labeled classes in the ontology. When no more exact matches can be found, it continues by stepping down the similarity threshold. In order to maintain scalability, it selects a small amount of random classes from the first ontology an then tries to find them in the second one.

Parameters

Parameter name	ValueType	Default	Description
PARAM_USE_GENERATOR	AlignmentGenerator	SimpleEvidenceMatcher(StringBasedSimilarity)	Defines the AlignmentGenerator to be used in the Hotspots.