Representação semântica de questionários de pesquisa por meio de ontologias

Abstract

Survey questionnaires, often represented in unstructured formats, present significant challenges regarding interoperability, reusability, and traceability. To mitigate these limitations, this thesis presents the Ontology for Questionnaires Sharing (OQS), a domain ontology based on the Unified Foundational Ontology (UFO) high-level ontology, designed to semantically represent questionnaires and their components, ensuring compliance with the FAIR (Findability, Accessibility, Interoperability, and Reusability) principles. The study adopted the Design Science Research (DSR) approach, and the development of the artifact was guided by the ontology engineering methodology Systematic Approach for Building Ontologies (SABiO). OQS not only formalizes essential elements of questionnaires, such as sections, questions, response alternatives, and variables, but also addresses frequently overlooked aspects, such as authorship, provenance, licensing, versioning, and branching. Each modeled entity receives a unique and persistent identifier, facilitating automated retrieval and effective reuse. The ontology was validated through use cases with widely used research instruments and by executing SPARQL queries, which satisfactorily answered the competency questions. In addition, an automated compliance analysis, performed with the FOOPS! Ontology Pitfall Scanner (FOOPS) tool, revealed that the OQS achieves 91% adherence to the FAIR principles, overcoming the limitations of related works identified in an integrative literature review. As a main contribution, the thesis makes an ontological artifact publicly available, which promotes transparency and reproducibility in open science. Future research can focus on integrating the OQS with data collection platforms and using artificial intelligence to automate the semantic annotation of legacy questionnaires in unstructured formats.