| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 382481 | Expert Systems with Applications | 2014 | 8 Pages |
•The work compares UML and SBVR as modeling languages for ontology development.•Comparison is performed by means of an exploratory experiment.•Feasibility of SBVR to OWL 2 mappings as ontology development technique is shown.•Effectiveness of UML for conceptualizing lightweight ontologies is shown.•Potential of SBVR to express complex notions of a domain of interest is shown.
Ontology development techniques still constitute an open research area despite its importance in semantic aware information systems. Until now, most methods have used UML in supporting ontology development process. Recent works propose the mapping of business rules expressions to ontology statements as a building technique by means of SBVR language. However, there is still no experimental research comparing such approaches.Aim of this work is to evaluate the feasibility of mapping business domain expressions to ontology statements. An exploratory experiment comparing performance of techniques based on UML and SBVR languages is presented. Comparison is rooted in the quality assessment of the ontologies developed by 10 equally sized groups randomly conformed by 30 undergraduate engineering students and applying such techniques.Developed ontologies largely outperform the minimally acceptable quality, according to the considered quality assessment framework. There is no statistical significant difference between the quality scores of the ontologies developed by means of UML and SBVR techniques, in any of the assessed quality dimensions.The feasibility of mapping business domain expressions to ontology statements is shown: ontologies developed by means of a SBVR based approach at least equate the quality of ontologies developed by using an UML based method. Results confirm previous research about the effectiveness of UML approaches for conceptualizing lightweight ontologies while stressing the potential of the SBVR language to express complex notions of a domain of interest. The potential of SBVR to OWL 2 mappings as an ontology development technique worthy of further study is highlighted.
