Résumé
Specialized ontologies are constructed to capture the skills of experienced experts in a particular domain, with the goal of sharing them with a larger community of trainees and less experienced experts in the domain. The main objective of this thesis is to construct a specialized ontology for the rising domain of simulation-based medical education, where formal models are lacking, and documentations are scarce. The thesis focuses on constructing an accurate and complete specialized ontology, and enriching and populating the constructed ontology.In this thesis, we have designed a four-staged collaborative ontology engineering methodology, which has resulted in the construction of the ontology called OntoSAMSEI. The first step is ontology bootstrapping (i.e., build a small initial ontology with the help of domain experts), followed by knowledge elicitation (fill the ontology using a questionnaire disseminated among the domain experts), enhancement (improve the core ontology by modeling commonalities), and update (enrichment and population). The resulting ontology is a hierarchy of classes and of properties, enriched by ontological constraints on the properties and on the classes.As a support to ontology update, we have designed and implemented a framework called IOPE for the automatic construction of a Graphical User Interface (GUI) consisting of pre-filled Web pages. The core idea behind IOPE is to transpose the RDF data and the ontological constraints into a GUI, using a set of mapping rules. These automatically generated GUIs provide guidance for domain experts and facilitate the ontology exploration and update through interactive graphical widgets. To finalize ontology updates, we propose a set of binding rules to specify how to transform user interactions into RDF graphs.The two contributions of this thesis are evaluated using an extensive and in-depth expert study, to show the benefits of IOPE and OntoSAMSEI in real-world use cases of medical experts.
Source: http://www.theses.fr/2021GRALM078
.
