The term ontology comes from ancient Greek and means something akin to “the doctrine of being”. It was originally a branch of theoretical philosophy or metaphysics and dealt with questions about the nature of being and the structure of reality. This focus shifted over time: in the more recent interpretations of philosophy, the term ontology concentrates primarily on language itself. This means that a linguistic division of areas of knowledge and life is carried out. In addition, the complex of ontology and language includes many principles. From this almost linguistic point of view, the term is also used in computer science.
Here, the creation of ontology serves to formalize knowledge and is constantly increasing in relevance. Because of the intensive use of the internet and the massive automation of various processes, it is more important than ever that different systems can communicate with each other. With the help of ontological systematizations, relevant information becomes readable and usable both for human users and for various machine systems (knowledge engineering).
When ontologies are created, not only concepts and terminologies themselves are captured, but also their relationships to each other and, if necessary, rules of derivation between them. In this respect, ontologies go far beyond pure linguistics in their application. They serve to facilitate communication between human and machine actors. In computer science, therefore, very specific ontologies are defined and used for each application or area.
Ontologies in computer science are characterized by three factors: A.) a formal structure so that machines can also read and use them. B.) The knowledge described by them must be unambiguously nameable. C.) In addition, ontologies describe not only concepts of a defined area, but also abstract concepts (“Specification of a Shared Conceptualisation”). These concepts are related to each other.
In order to describe the structure of an ontology, one has to take a closer look at the concepts and relations of ontologies. There are two types of concepts which can be differentiated from one another: those that describe an entire class or a specific set of individual objects, and concepts that exclusively refer to concrete terms.
The relations are divided in two: There are relations which arrange concepts hierarchically or which put concepts into arbitrary relationships with one another.
The semantic search is one of the main applications in the semantic web – especially for unclear or unknown knowledge. Ontologies can be useful here in several places at once. On the search query side, the input can be generalized or specialized, as well as adjusted or corrected in terms of content if the information searched for is based on ontological structural knowledge.
In search processing, ontological derivation knowledge and ontological definitions serve to bridge possible contradictions in the formulation of the search and the available information. Furthermore, a similarity-based search is made possible by consulting background knowledge. If documents are enriched by further descriptive (and ontologically-structured) meta data, even complex contents can be displayed in the search.
Sorting and presentation rules formulated by ontologies can also be applied to search results so that information can be optimally displayed and integrated.
This leads to the next frequent area of application for these classification systems: intelligent information integration. Ontologies can also help to describe not only content, but also schemata. This is the case, for example, with the integration of data that has different origins (sources). Ontologies function as transformation and translation standards.