The Use of Large Language Models for the Analysis of Professional Competencies in the Regional Labor Market of the Republic of Belarus

An integrated approach to the analysis of professional competencies in the Belarusian labor market using large language models is presented. A methodology is proposed that includes data collection using web scra­ pers, preliminary processing using a multi-level cleaning system and normalization of text information, classification and analysis of competencies based on interaction with large language models. Detailed skill profiles for each professional group are formed, clusters of complementary competencies are identified, and stable combinations of skills required in various professional fields are determined. High efficiency of using large language models for the tasks of extracting information on competencies from unstructured text descriptions of vacancies is demonstra­ ted with accuracy and completeness indicators exceeding 85%. A methodology for labor market analysis has been developed that integrates traditional big data analysis methods with the capabilities of modern language models.