An Approach to Evaluating Academic Courses Through Multi-Factor Metrics and Reference Course Alignment (In Russ.)

This paper presents a system of metrics for the automated evaluation of university-level academic courses, based on the analysis of topic formulations and their structural organization. The objective of the study is to develop a formal-ized quantitative methodology for assessing course quality, accounting for both internal content characteristics and alignment with a reference syllabus (when available). The proposed approach includes reference-independent metrics that assess logical coherence, structural balance, and content redundancy, as well as reference-based metrics that eval-uate topic coverage and thematic relevance. Analytical tools include graph-based representations, semantic embeddings, and sequence-based metrics. The methodology was tested on a corpus of mathematics-related university cours-es. The results demonstrate the system’s effectiveness in identifying logical inconsistencies, missing core topics, and irrelevant content. The proposed approach can be used to support the design and revision of academic courses and to provide automated feedback in digital learning environments. As a next step, the integration of the system with generative language models and the extension of the metric set to include content-level characteristics—such as topic depth and complexity levels—is being considered.