Pr4 UG co Ontology

From Protege Wiki
Revision as of 13:55, July 27, 2010 by RichMorin (talk | contribs) (Automated import of articles)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Ontology

comment   Editing   Glossary   Outline


Concepts
Annotation
Ontology

An ontology is a formally-defined vocabulary for a particular domain of interest. Because it is formally defined, it can reduce the opportunity for confusion in human discourse. Indeed, some ontologies are created for this exact purpose.

However, ontologies are typically created to enable a computer program (eg, a reasoner) to infer logical consequences about the domain. So, developing an ontology might be a precursor to creating a database, expert system, or knowledge base.

The restriction to "a particular domain of interest" may seem odd. However, there are both practical and theoretical reasons for limiting the subject matter and the encoded information. In brief, the problem is that the meaning of a term depends on the kinds of questions being asked.

For example, even something as prosaic as an onion will be seen in different ways to a biologist, a cook, a farmer, and a grocer. Details which are critical to one person may be completely irrelevant to another.

Structure

An OWL ontology (as used in Protege 4) is based on a class hierarchy. Each class defines a level of abstraction. For example, the Pizza Ontology defines classes such as Pizza, VegetarianPizza, and Margherita.

The hierarchy is then "decorated" with properties. Each property specifies a binary (ie, dyadic, two-ended) relationship. A relationship might indicate that a Pizza has:

  • a PizzaBase,
  • one or more PizzaToppings, or
  • certain standardized sizes.

Because of inheritance, Pizza's descendant classes (eg, VegetarianPizza, Margherita) also have these properties.