Abstract
In the context of rapidly evolving digital economy requirements, the intellectual support of students’ professional development based on the analysis of educational data and labor market demands has become an important challenge for higher education institutions. The aim of this study is to develop a methodology for the intelligent analysis of educational data and the architecture of an analytical system to support the professional development of students in technical disciplines using Educational Data Mining (EDM), Learning Analytics (LA), and Large Language Models (LLMs). The proposed approach integrates educational programs, students’ academic data, digital portfolios, and employer requirements obtained through the APIs of the HeadHunter and Enbek platforms. The proposed methodology comprises a sequence of stages, including the collection of educational data, integration of heterogeneous information sources, preprocessing, cleaning and structuring of textual documents, automated extraction of competencies, semantic matching of these competencies with labor market requirements, and the generation of analytical recommendations. The proposed solution is based on the combined use of Educational Data Mining techniques, semantic competency matching using GPT-4 (OpenAI API), and intelligent gap analysis. The developed architecture includes modules for the automated extraction and structuring of data from PDF and DOCX documents, intelligent competency analysis, prediction of students’ professional readiness, and generation of personalized recommendations for educational pathways. The preprocessing stage includes text extraction from PDF and DOCX documents, removal of technical elements, normalization of document structure, standardization of competency representation formats, transformation of the data into a structured JSON format, and subsequent semantic analysis using GPT-4 (OpenAI API). The experimental evaluation was conducted using 32 academic programs, 1,247 IT-related job vacancies, and data from 286 engineering and technical students. The obtained results demonstrate the practical feasibility of the proposed educational data mining methodology and confirm the possibility of intelligent matching between educational competencies and labor market requirements based on Large Language Models (LLMs).