Glossary · Semantic interoperability and information models
Ontology alignment
Also known as: Ontology matching
German: Ontologieabgleich
In semantic technology, ontology alignment is the process, and its result, of identifying correspondences between the concepts, properties and relations of two or more ontologies, such as equivalence, subsumption or relatedness, so that data described with them can be linked or integrated.
- Information models
In one sentence
Ontology alignment identifies correspondences, such as equivalence or subsumption, between the concepts of two or more ontologies.
Example
A project finds that the class Motor in the plant ontology is equivalent to an ECLASS class and narrower than Electrical machine in a standard ontology.
How it applies
- Engineering: Alignment is needed when companies or tools use different ontologies or vocabularies for the same domain. It can be done manually by experts, supported by automated matching tools.
- Quality: Automated matching based on names alone is error-prone; similar labels can hide different meanings. Experts should review proposed correspondences, especially equivalences.
- Documentation: Document each correspondence with its type, rationale, author and date. Alignments need maintenance when either ontology changes, otherwise Semantic drift creeps in.
Ontology alignment vs. ontology mapping
Alignment establishes which concepts correspond and how. Ontology mapping expresses these correspondences in an executable or formal form, for example as transformation rules, so that data can be translated from one ontology to the other. Some authors use both terms synonymously.