A hybrid Bayesian BWM-machine learning framework for university digital transformation assessment: Integrating expert clustering and predictive validation

The rapid advancement of artificial intelligence–driven technologies has underscored the need for universities to undergo a comprehensive digital transformation. This study introduces a novel hybrid framework that combines the Bayesian best–worst method (BBWM) with machine learning techniques to identify and assess key success factors for university digital transformation. This study integrates expert clustering analysis and predictive modeling validation to provide enhanced decision support capabilities. Through systematic evaluation of 27 experts across 5 key dimensions—digital infrastructure, teaching and learning, research, administration and governance, and stakeholder engagement—the analysis revealed critical insights into transformation priorities. Clustering analysis identified seven distinct stakeholder assessment patterns, demonstrating significant heterogeneity in evaluation approaches across professional backgrounds and experience levels. Digital skills and literacy development emerged as the most influential factor, followed by the availability of online learning resources and the adequacy of network infrastructure. Methodological validation demonstrated exceptional convergence between the BBWM and traditional BWM (correlation coefficient: 0.9961), whereas machine learning validation confirmed the robustness of dimensional priority hierarchies through feature importance analysis. The hybrid framework achieved a ranking confidence of 70.23 % and provides stakeholder-specific insights for developing differentiated implementation strategies. This study contributes to the digital transformation literature by presenting the first integration of Bayesian inference and machine learning for university assessment, providing evidence-based frameworks to support strategic technology planning and resource allocation in an increasingly digital educational landscape.