The success of using ontologies to solve knowledge-related or semantic interoperability problems is related to the quality of the used ontologies. The quality of an ontology, in turn, is strongly related to the quality of the languages, methods and tools used to develop it. Thus, it is important to advance on the theoretical and practical support for ontology engineering. In this sense, we have proposed several methods, languages and tools to aid in ontology engineering.
SABiO (Systematic Approach for Building Ontologies)
SABiO was first proposed in 1997. It was used over the years to develop several ontologies in different domains, such as Software Engineering and Cardiology. Lessons learnt from the experiences using SABiO and the analysis of its strengths and weakness led the method to be evolved. In 2014, the current version of SABiO was published. SABiO supports ontology development by providing a set of processes and activities to be followed to produce reference ontologies (i.e., conceptual models built with the aim of describing the domain in reality, without any concern regarding computational properties) and operational ontologies (i.e., ontologies built for machine interpretation). According to SABiO, when developing an ontology, the ontology engineer should consider the Development Process plus five support processes, namely: Knowledge Acquisition, Documentation, Configuration Management, Evaluation, and Reuse. The current version of SABiO is presented in:
EArly-OE (Enterprise Architecture-driven early Ontology Engineering)
EArly-OE establishes strategies for the use of EA (Enterprise Architecture) models as non-ontological resources to provide knowledge in initial ontology engineering activities. It is a particularly suitable method for developing ontologies in domains rich in structured processes. EArly-OE prescribes guidelines for the use of elements of EA models to support initial activities of ontology engineering, including: identification of domain specialists and potential ontology users; selection of consolidated knowledge resources in the domain of interest; definition of the ontology intended uses; identification of the ontology scope and elicitation of functional requirements; and initial proposal for ontology modularization. Early-OE is described in:
CLeAR (Conducting Literature Search for Artifact Reuse)
CLeAR is a systematic approach to find and select reusable knowledge resources for building ontologies with the purpose of scientific research data integration. It follows some principles of Systematic Literature Review, supporting the search for knowledge resources in the scientific literature. CLeAR includes activities organized in three cycles. The first cycle aims at defining the data integration requirements and the scope of the ontology to be developed. The second cycle aims at systematically identifying structured resources candidates to be reused in the development of the ontology, based on the requirements defined in the first cycle. In the last cycle, structured resources are selected to be reused. CLeAR addresses specific ontology engineering activities. As a consequence, it was designed to be used as a complement to existing ontology engineering methods, such as SABiO (see above). CLeAR was proposed in:
GO-FOR (Goal-Oriented Framework for Ontology Reuse)
GO-FOR applies GORE (Goal-Oriented Requirement Engineering) in Ontology Engineering to express the design rationale of ontology model fragments. In GO-FOR, ontology models are depicted in fragments (i.e., domain ontology patterns) related to goals. These model fragments are self-contained ontology structures called Goal-Oriented Ontology Patterns (GOOP), a new type of pattern to be applied to develop ontologies in a goal-oriented approach. In GO-FOR, goals can be used as parameters to support ontology shareability and reuse. In addition to GOOPs, GO-FOR introduces GOOPR (GOOP Repository), a repository to store GOOPs and that serves as an abstraction layer for ontology development. To support GO-FOR use, we developed GOOP-Hub, a tool that supports GOOPs creation, search and retrieval. An overview of GO-FOR is presented in:
Guidelines on how to use GOOP-Hub are available in the GOOP-Hub User Guide.
Integra is an approach for ontology development based on integration, which uses goal modeling to support requirements elicitation activities, help make explicit design rationale and aid in the search for candidate ontologies for reuse. Integra prescribes a process composed of four phases: Ontology Requirements Elicitation, Selection of the Ontologies to be Integrated, Ontology Integration, and Evaluation of the Resulting Ontology. The main result of the application of Integra is a reference ontology (i.e., a conceptual model built with the aim of describing the domain in reality, without any concern regarding computational properties). In case there is a need for an operational ontology (i.e. an ontology built for machine interpretation), then a design phase should be executed to adjust the reference ontology for implementation. Both design and implementation are out of the scope of Integra. They are addressed in other ontology engineering methods, such as SABiO (see above). Hence, Integra may be combined with other methods for designing and implementation of operational ontologies. Integra was proposed in:
The specification of Integra, containing a detailed description of its phases and activities, and a practical example of Integra use will be available here soon.