Reúso de recursos da web semântica para a construção de vocabulários controlados no contexto da ciência da informação

Abstract

This thesis addresses as a subject the reuse of Semantic Web resources for construction of instruments of knowledge representation, framed in the area of Library and Information Science (LIS). The aim of this research was to explore the literature that covers the areas of LIS (Library and Information Science) and Computer Science (CS) regarding the construction of controlled vocabularies (CV). The literature review has shown that reuse is recognized as an important step in CV construction. With the reuse of resources, time and effort are saved, instead of starting the construction from scratch and besides, it promotes the interoperability between agents (humans and machines). The methodologies of construction of CV that are mentioned in this thesis cite reuse as an important task and that must always be present in the process of building Knowledge Organization Systems (KOS). In the field of LIS there are few methods that explicitly recommend the reuse of resources provided by the semantic web. In addition to the exploration of issues related to knowledge representation, a guide was proposed to assist LIS professionals in the creation of knowledge representation tools, which was named OntoM4IS+ (method of reusing ontological and non-ontological resources for information science). It was based on several methodologies as well as good systems modeling practices, promoting data description in order to facilitate interoperability between agents, thus providing future reuse. By its nature, this research is applied, that is, that research that does not take into account only the fundamental understanding, which is proper to basic science, but which is concerned with considerations of use. Regarding to the objectives, the research is exploratory, and as approach to problem, could be considered qualitative. The research method adopted was Design Science Research (DSR), a method that has a qualitative approach and is part of the spectrum of Pragmatism. Regarding data collection procedures, firstly was defined the keywords that were used for document retrieval in the main national databases (CAPES’s Journals Portal) and international databases (Web of Science, RCAAP, Scopus, NDLTD) were first defined. Subsequently, other data processing techniques such as the reading grid were combined and in addition a matrix of concepts was created for all retrieved documents. The theoretical framework was based on the theories and practices of LIS regarding the organization of knowledge and also using the technologies and standards of the Semantic Web for data representation, from which stand out among others the Resource Description Framework (RDF) and the Web Ontology Language (OWL). Also was mentioned the various formats of serialization such as RDF/XML and Turtle, always following the principles of open data (Linked Data). The evaluation of OntoM4IS+ was performed in an iterative and incremental manner. In the first phase, it consisted of the submission of articles to events and peer review scientific journals. Contributions were also received from a meeting with one of the most renowned international experts in the field of knowledge organization, Dagobert Soergel, professor at the University at Buffalo/State University of New York. At a later stage, OntoM4IS+ was evaluated in the experimental situation with the evaluation of Embrapa's OntoAgroHidro domain ontology, in light of what is established by OntoM4IS+. As a result, it is believed that in addition to the research artifact, the work contributed to bring to BCI a research method that is still relatively little used but that fits the nature of research in BCI, which is the DSR.