Ontology
| Collection | GlossariumBITri |
|---|---|
| Author | Jorge Morato-Lara |
| Editor | Sonia Sánchez-Cuadrado |
| Year | 2010 |
| Volume | 1 |
| Number | 1 |
| ID | 71 |
| Object type | Concept Resource |
| Domain | Artificial Intelligence Semantic Web |
| es | ontología |
| fr | onmtologie |
| de | ontologie |
According to Gomez-Perez (2004), Ontology definition has evolved during the last twenty years. In 1995, Guarino collects seven definitions about this concept to propose a new one. This author defines ontology as “a set of logical axioms designed to account for the intended meaning of a vocabulary” (Guarino, 1998).
One of the reasons for the disagreement is the broad definition that has been proposed to group all current ontologies. Wikipedia defines ontology as a “Formal representation of a set of concepts within a domain and the relationships between those concepts”. These definitions might be right even for any other Knowledge Organization Systems or terminological resource. Wikipedia adds that ontologies are “used to reason about the properties of that domain, and may be used to define the domain”.
The best known definition was proposed by Gruber: “a formal, explicit specification of a shared ontology.” (1991).
Modeling techniques
Two are the most commonly used techniques:
- First-order logic
- Description logic
Ontology elements
Depending on the technique that has been used, the vocabulary to design some elements might be different.
- Classes: It is a set of similar individuals. These sets represent the main concepts of the domain. These concepts are often arranged in a hierarchical way. Classes might have attributes and functions and can be linked to another class by relations.
- Relations: Relationships to link classes and individuals
- Attributes: properties or slots those classes and its individuals can have.
- Functions
- Individuals: instances or objects of a class.
To perform inference, the existence of assertions considered true is needed: these assertions are used to express restrictions, rules and axioms.
Finally, the Events are a way to represent how the value of attributes and the relationships might change.
First order usually calls these elements: classes, relations, attributes (slots), functions, instances, and axioms.
Description Logic uses the following elements: concepts (equivalent to classes); roles (equivalent to relations and properties of concepts); and Individuals (equivalent to the instances of concepts and their properties).
Ontologies principles
To be able to share knowledge, interoperability is required. Many principles have been proposed (Gruber 1993):
- Clarity: objective definitions, formalized with axioms, and complete (necessary and sufficient conditions).
- Minimal Encoding Bias
- Extendibility
- Minimal ontological commitments
Gomez-Perez (2004) adds to this list:
- Representing disjoints and exhaustive knowledge
- Minimizing distance between siblings
- Standardizing names in a clear form
Types of Ontologies
There are different types of ontologies:
Upper ~ (top ~ or foundational ~): describe very general concepts that are common to all the ontologies. Other ontologies can be aligned with these concepts by their root term. Examples are DOLCE, Proton, SUMO, and CYC.
Task ~: describe the vocabulary related to some generic task or activity.
Domain ~: concepts of a domain and their relationships.
Common ~ (generic ontologies): common knowledge reusable in different domains. Examples are ontologies about time or space.
Knowledge Representation ~: primitives to express knowledge in a formalized way.
Application ~: it is an ontology adapted to a specific application.
Regarding the reusabilitity and usability, more abstract ontologies are highly reusable (Knowledge Representation Ontologies and Upper ontologies) but their usability is poor. Application and Domain Ontologies have a low reusability but a high level of usability.
Languages
Modeling choices are attached to different languages. As an example, one of the languages related to first order logic is KIF; OWL is usually related to frame logic. OWL is a language widely used to represent ontologies in the Web. OWL serialization is based on RDF/RDFS.
References
- GUARINO, N. (1998) “Formal Ontology in Information Systems”. In: GUARINO N (ed). FOIS 98. Trento: IOS Press. Pp. 3-15
- GRUBER, T.R. (1993) “A translation approach to portable ontologies”. Knowledge Acquisition, 5(2):199-220, 1993
- GRUBER, T.R. (1993) “Toward principles for the design of ontologies used for knowledge sharing”. Int. Workshop on Formal Ontology in Conceptual Analysis. Padova 1993.
- GOMEZ-PEREZ, A.; FERNANDEZ-LOPEZ, M.; CORCHO, O. (2004) Ontological engineering: with examples from the areas of knowleddge management, e-commerce and the Semantic Web. London: Springer.