Ontology: A Web of Meaning


Tom Gruber was affiliated with Stanford University when he penned this oftcited definition of an ontology:

Ontology

"An ontology is a specification of a conceptualization."[15]

Gruber points out that ontology has a rich history in philosophy, where it deals with the subject of existence and is often associated with epistemology. His definition, originally for the artificial intelligence community, has to do with concepts and relationships, and in particular how agents (computer programs) can commit to sharing definitional information.

Typically ontologies are "graphs" or networks, rather than hierarchies, although this isn't a requirement. More important, ontologies have a much richer set of relationships, constraints, and rules, such that we can reason about the information in the ontology, and, by extension, reason about items that have been classified by the ontology.

Medicine has been a very rich field for the development of ontologies. UMLS is a large but not extremely rigorous ontology of medical concepts. Its primary use is to help with searches, so having a large graph of interrelated concepts is helpful rather than cumbersome. The UMLS ontology has several almost orthogonal taxonomies for categories such as "anatomy" (the thigh bone's connected to the knee bone), "system function" (circulatory, digestive, etc.), and "morphology" (shape; e.g., distended, elongated, etc.). The UMLS ontologies contain rich sets of cross-taxonomy relationships, such as "the heart valve is a part of the heart," "the heart is a member of the circulatory system," "the heart is in the thorax," and "the heart is above the diaphragm."

WordNet is a public domain lexical database. It serves as a base for many ontology projects that deal with natural language and lexical issues; however, it is not an ontology in itself.[16]

Ontologies are now relatively easy to build, and thousands of them are springing up. This isn't a problem, except that there is a desire to have ontologies use each other to extend their range of knowledge. One factor that will greatly accelerate this is the creation and acceptance of what has been called an "upper ontology."[17] The basic idea of an upper ontology is that it would be a base ontology that could be shared across many domains. In other words, if other ontologies were expressed in terms of the upper ontology, an agent would only need to know the upper ontology to be able to work out most if not all of a subscribing ontology.

[15]Tom Gruber, "A Translation Approach to Portable Ontologies," Knowledge Acquisition, vol 2, pp. 199–220, 1993.

[16]The WordNet database can be downloaded at http://www.cogsci.princeton.edu/~wn/.

[17]IEEE has a working group at http://suo.ieee.org/.




Semantics in Business Systems(c) The Savvy Manager's Guide
Semantics in Business Systems: The Savvy Managers Guide (The Savvy Managers Guides)
ISBN: 1558609172
EAN: 2147483647
Year: 2005
Pages: 184
Authors: Dave McComb

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net