Expressibility of OWL Axioms with Patterns

Aaron Eberhart, Cogan M Shimizu, Sulogna Chowdhury, Md Kamruzzaman Sarker, Pascal Hitzler

Keywords: Ontology, OWL Axioms, Semantic Expressibility, Simple Axioms, Axiom Patterns

Abstract:

The high expressivity of the Web Ontology Language (OWL) makes it possible to describe complex relationships between classes, roles, and individuals in an ontology. However, this high expressivity can be an obstacle to correct usage and wide adoption. Past attempts to ameliorate this have included the development of specific, presumably human-friendly syntaxes, such as the Manchester syntax or graphical interfaces for OWL axioms, albeit with limited success. If modelers want to develop suitable OWL axioms it is important to make this as easy as possible.

In this paper, we adopt an idea from the Protege plug-in, OWLAx, which provides a simple, clickable interface to automatically input axioms of a limited number of types by following simple axiom patterns. In particular, each of these axiom patterns contains at most three classes or roles. We hypothesize that most of the axioms in existing ontologies could be expressed semantically in terms of simple patterns like these, which would mean that more complex patterns can be used very sparingly.

Our findings, based on an analysis of 518 ontologies from six public ontology repositories, confirm this hypothesis: Over 90% of class axioms in the average ontology are indeed expressible with our simple patterns. We provide a detailed analysis of our findings.

Subtrack: Semantic Data Management, Querying and Distributed Data

First Author Is Student: Yes