Abstract
The article analyses the influence of digital transformation on the economic development of European countries using a combination of classical econometric approaches and machine learning algorithms. The study uses 85 indicators of digital economy and society for 27 countries during 2017–2022, covering various aspects of digitalisation: human capital, digital infrastructure, broadband coverage, ICT specialisation, business innovation activity, etc. After preliminary data processing, multicollinearity diagnostics, and hierarchical clustering, factor analysis identified four latent components: digital competence and business innovation, digital infrastructure and connectivity, broadband coverage and penetration, ICT human resources and specialisation. To evaluate the relationships between digital factors and GDP per capita, pooled OLS, ridge, lasso regressions, random forest, XGBoost, and support vector regression models were applied. The highest forecasting accuracy was demonstrated by the SVR model, which provided minimal error values and effectively captured nonlinear dependencies in panel data. Feature-importance analysis revealed the leading role of digital competence and business innovation, as well as the considerable cross-country heterogeneity of digital drivers of economic development. The results confirm the need for developing differentiated digital-policy strategies and provide the basis for further advancement of causal and spatial modelling of the digital economy.