The MOUSE approach : mapping ontologies using UML for system engineers

Abstract

To address the problem of semantic heterogeneity, there has been a large body of research directed toward the study of semantic mapping technologies. Although various semantic mapping technologies have been investigated, facilitating the process for domain experts to perform a semantic data integration task is still not easy. This is because one is required not only to possess domain expertise but also to have a good understanding of knowledge engineering. This thesis proposes an approach that automatically transforms an abstract semantic mapping syntax into a concrete executable mapping syntax, I call this approach MOUSE (Mapping Ontologies using UML for System Engineers). In order to evaluate MOUSE, an implementation of this approach for a semantic data integration use case has been developed (called SDI, Semantic Data Integration). The aim is to enable domain experts, particularly system engineers, to undertake mappings using a technology that they are familiar with (UML), while ensuring the created mappings are accurate and the approach is easy to use. The proposed UML-based abstract mapping syntax is evaluated through usability experiments conducted in a lab environment by participants who have skills equivalent to real life system engineers using the SDI tool. Results from the evaluations show that the participants could correctly undertake the semantic data integration task using the MOUSE approach while maintaining accuracy and usability (in terms of ease of use).