Theoretical and Methodological Approaches to Analysis and Forecasting of the Labour Market within the Framework of Sustainable Development

Abstract

The escalating risk of precarious employment underscores the necessity to examine the impact of digitalization of the labour market on sustainable economic development. The precarization of employment may result in diminished performance metrics for the achievement of decent work goals. The study is aimed at the development of a theoretical approach in order to supplement sustainable development indicators with the characteristics of vacancies in the labour market in the context of its digitalization. Official statistical data, as well as online information sources for the period 2013-2023, summarized with the help of Big Data and AI technologies serve as the empirical basis of the study. A forecasting model for demand metrics of the labour market was chosen on the basis of evidence from the labour market in Minsk. The XGBoost model demonstrated superior results compared to other time series models. The results of October 2023 - October 2024 forecasting of the number of vacancies in Minsk are given in the article. The results are received after the predictive model of gradient boosting was adjusted. The predictive model of gradient boosting was trained using actual data. Further research to improve the quality of forecasts will focus on the following: optimization of the model of gradient boosting using deep learning; identification of additional features and factors that influence appearance of new vacancies in the regions of Belarus.