Student: Jose Illescas Martinez     (2020)

Supervisor: a.Univ.-Prof. DI Dr. Wolfram Wöß

Motivation and Challenges

An ontology is a formal, explicit specification of a shared conceptualization that is characterized by high semantic expressiveness required for increased complexity.

Although the use of ontologies has become increasingly relevant in many research projects, their adoption to business applications still lacks behind. One obstacle is the absence of a simple graphical notation that supports the conceptualization of an ontology.

Existing tools, such as Protégé, assist ontology engineering with a step-by-step creation process of a machine-readable formal representation of an ontology. However, in practice important decisions concerning the concepts and structure of an information system are made in a design phase previous to the ontology development. These early decision steps are especially important for ontologies, because they enable the collaboration of a number of different stakeholders (cf. sharing principle), who have to find an agreement on common concepts. A standardized and simple graphical notation as it is available for conceptual data modeling (e.g., entity-relationship model, UML class diagram) would support the communication between technicians and non-technicians and would provide graphical ontological models which are useful and important for documentation purposes as well as a common understanding in negotiation and design meetings.

In addition, ontology development and administration tools would increase their user-friendliness by offering visualization features.
(1) Forward-engineering: The user develops an ontology model, supported by a graphical user interface and the standardized notation, which can be automatically transformed to a formal ontology (e.g., OWL, RDF, Manchester syntax).
(2) Reverse-engineering: Existing formal ontologies can be visualized with the standardized notation.

Objective

The main objective of this master's thesis is the development of a graphical notation for ontological engineering. The notation should include representations for concepts (classes), data type properties (similar to attributes of an entity), object properties (similar to relations between entities, but with an explicit direction for determining domain and range), and class axioms/restrictions. In addition, simple rules for transforming an ontology model into a formal ontology language (such as OWL DL, RDF, Manchester syntax) should be proposed. These simple rules can be similar to transformation rules required for transforming an entity-relationship model into a relational model used in the field of relational databases.
The transformation rules can be evaluated with a reverse-engineering implementation that takes a formal ontology as input and visualizes it with the developed notation.