Konzeption und Entwicklung eines Systems zur Bewertung und Beurteilung von OWL Modellen im Rahmen des intelligenten tutoriellen Systems eTutor

The Institute for Data & Knowledge Engineering at the JKU uses the Intelligent Tutoring System (ITS) eTutor to enrich their teaching. This system uses expert modules to grade assignments automatically, with each module covering only a small area of expertise. No such module exists for the field of ontology design, so the assignments handed in by the students still have to be graded manually. Solving this problem is the purpose of this thesis. After comparing different approaches a new concept has been developed, which combines methods for aligning ontologies with findings from teaching OWL-DL at the University of Manchester. This enables the detection of mistakes made by students regarding their instructions, formal logic and OWL-specific issues. The concepts defined in the submitted ontology are being analyzed in the presence of a model solution. Subsequently, conclusions are drawn by combining the results of this analysis. The resulting feedback allows students to improve their submissions. Based on this concept, an expert module has been implemented in Java. This software can either be used stand-alone or within the ITS eTutor. It uses a modular approach and can easily be extended to detect mistakes in all kinds of entities available in OWL 2.