OntoMI: Uma Ontologia Fundamentada na Teoria das Inteligên cias Múltiplas para Classificação Semântica de Recursos Educacionais

Abstract

The formal and explainable representation of educational theories in computational systems remains a challenge for teaching-personalization initiatives, especially when the goal is to trace inferences back to observable textual evidence. In this context, this study investigates the potential of an ontological and computational approach to classify and semantically represent textual educational resources, adopting the Theory of Multiple Intelligences as the pedagogical reference for instantiation. The research was conducted under the Design Science Research paradigm, integrating a systematic literature review, ontology engineering methods through Ontology Development 101 and SABiOx, and a proof of concept. As results, a pedagogical practices guide, the OntoMI ontology, and the Intelli3 computational model were produced. OntoMI structured the link between linguistic evidence and inferred categories, promoting traceability and support for explainability. Intelli3 generated vector-based profiles from the instantiated knowledge and enabled the ranking of instructional materials according to multiple-intelligence profiles. Empirical tests indicated stability and precision in differentiating profiles, supporting the technical feasibility of semantic classification when computational inference is grounded in explicitly stated evidence. It is concluded that the ontology serves as the formalization core and mediation layer between pedagogical theory and computational operationalization, supporting teacher decision-making without replacing the educator’s interpretive role.