Abstract

Current ontology engineering methodologies and tools follow an iterative ontology authoring approach where Knowledge Engineers gather the requirements for the ontology from the Domain Experts interpret them and add them to the ontology. Domain Experts can verify, validate or suggest revisions in the ontology. The recent increase in popularity of lightweight OWL (Web Ontology Language) ontologies has created the need for faster, more agile and effective methods to support the task of ontology creation. This need is testified by efforts that focus on enabling Domain Experts to better participate in the process of ontology construction, making them able not only to provide domain knowledge to Knowledge Engineer but also to directly contribute to the editing of certain parts of the ontology. The contribution of this dissertation is threefold. In the first part of this dissertation, we investigate how to reuse already formalised knowledge, such as foundational, core, and domain specific ontologies, to support the construction of templates for the enterprise modelling domain. Templates are a popular strategy for involving Domain Experts in ontology authoring but a typical problem with this approach is that Knowledge Engineers have to create them from scratch every time they face a new modelling scenario. Libraries of reusable templates can help the work of Knowledge Engineers and support Domain Experts in providing their domain knowledge. We report our findings and lesson learned both from the concrete experience of building the templates and from a survey carried out for their evaluation. The second contribution of this dissertation is the reuse of already formalised knowledge to enrich domain specific taxonomies with knowledge involving relations by means of restriction axioms. In this contribution, at first we have presented a skeleton of a very general method called GENERATOR (Guided entity reuse and class expression generator). Second, a specific instantiation of GENERATOR, called FORZA (Foundational Ontology and Reasoner-enhanced axiomatiZAtion), for the reuse of knowledge contained DOLCE-lite and mereoTopoD is defined. As a final contribution of this dissertation, we present an approach based on Decision Diagrams or a Question-Answering (QA) system to choose the most appropriate category from a reference model (ontology or reference schema) without being exposed to the model itself. This contribution is instantiated in two different Decision Diagrams. The first one for the selection of the most appropriate template among the ones developed for the enterprise modelling scenario, and the second one for choosing the most appropriate DOLCE-lite category to support the FORZA method. As a proof-of-concept, implementation of FORZA and templates has been integrated with the MoKi ontology development environment.