A method for formalizing educational texts

Abstract

Context and relevance. The formalization of educational texts is necessary for content analysis, checking the logic of exposition, and developing computer-based support for learning; however, arbitrary verbal texts are difficult to formalize automatically without taking their genre and functional features into account. Objective. The study aims to develop a method for transforming an educational text into a formal model that preserves the content and sequence of presentation of the educational material as fully as possible. Hypothesis. If an educational text is considered as a regulatory text consisting of a sequence of denotative situations, each sentence can be represented through a predicate, actants, circumstants, and the results of the described action, while relations between sentences can be represented as edges of a directed graph. Methods and materials. The study draws on semantic syntax, the concept of proposition, frame structures, and graph-based knowledge representation; the material is a verbal educational text considered as an object of content formalization. Results. An algorithm is proposed in which sentences are sequentially transformed into groups of frame instances that form subgraphs and are then combined into a single graph based on matching content components and the logic of their introduction. Conclusions. The method provides a basis for constructing ontologies of subject-domain concepts, identifying coherence violations, and developing computer-based tools to support assessment of learning outcomes and learners' self-study.