Abstract:
Across a wide range of domains, there is an urgent need for a wellfounded
approach to incorporating uncertain and incomplete knowledge into formal
domain ontologies. Although this subject is receiving increasing attention
from ontology researchers, there is as yet no broad consensus on the definition of a
probabilistic ontology and on the most suitable approach to extending current ontology
languages to support uncertainty. This paper presents two contributions to
developing a coherent framework for probabilistic ontologies: (1) a formal definition
of a probabilistic ontology, and (2) an extension of the OWL Web Ontology
Language that is consistent with our formal definition. This extension, PR-OWL,
is based on Multi-Entity Bayesian Networks (MEBN), a first-order Bayesian logic
that unifies Bayesian probability with First-Order Logic. As such, PR-OWL combines
the full representation power of OWL with the flexibility and inferential
power of Bayesian logic.
