Options
2021
Conference Paper
Titel
Ontology Design Patterns for Representing Context in Ontologies using Aspect Orientation
Abstract
Context-sensitive knowledge is ubiquitous. Knowledge may, for example, change over time and from location to location, or it may be dependent on an observer's background knowledge, belief, or opinion. We observe that specifying context leads to recurring modeling problems. For example, the modeling of temporal context involves the specification of one or several time intervals and attaching them to the axioms or facts that are supposed to be valid during that time. The formalism for specifying the temporal context may, however, vary. As a solution, we present a catalog of parametrizable ontology design patterns (ODPs) specific to the problem of modeling context. We base the patterns on the aspect-oriented extension of OWL 2, because it allows nesting of contexts and the usage of the entirety of OWL 2 (DL) as the context description language. We evaluate the adequacy and usefulness of the approach within a real-life research project about anatomical structure recognition in 3D endoscopic imaging, in which a highly contextualized mapping ontology between the qualitative arrangement of 3D shapes, anatomical structures, and surgical situations is developed. We can show that the use of a context specification formalism leads to an adequate representation of the domain at hand and that the proposed context ODPs facilitate the modeling of contextualized knowledge. Nevertheless, the ODPs are sufficiently abstract and general to be reused in different domains and application scenarios.
Author(s)