STRUCTURAL AND MATHEMATICAL MODEL OF A RECOMMENDATION SYSTEM FOR THE SELECTION OF EDUCATIONAL RESOURCES BASED ON ONTOLOGICAL AND SEMANTIC CONSTRAINTS

Abstract

The paper proposes a structural and mathematical model of a recommendation system for the selection of educational resources based on the integration of metadata from academic disciplines and library collections. The model formalizes the key elements of the educational process in the form of object sets and a system of relations between them and introduces a comprehensive relevance function that combines semantic similarity of textual descriptions, ontological distance between classes, and contextual educational constraints.

The semantic component is implemented through vector representations of annotations and keywords using natural language processing techniques. The ontological component is based on a formalized hierarchy of the subject domain, while the contextual component takes into account the level of education and the year of study.

The proposed approach restricts the search space, increases the interpretability of recommendations, and ensures scalability as the volume of data grows. The resulting model provides a formal foundation for the implementation of an intelligent recommendation platform within the digital educational environment of a higher education institution.