Enhancing infrastructure decisions through meta-variable analysis and I-BIM via LightGBM

In the post-COVID-19 era, economic uncertainty and restricted budgets have made countries more cautious about investing in construction. While building information modelling (BIM) has transformed building projects, its application to infrastructure remains difficult. This study investigates whether BIM’s long-standing philosophy can be effectively adapted for infrastructure through a tri-axial model based on the policy-technology-process framework. A competency structure was developed, including five core meta categories and 26 sub-variables. A structured questionnaire was distributed in the USA and Turkey, chosen for their contrasting BIM maturity levels. Domain competency analysis introduced the total number of meta-variables factor at three levels to assess readiness. Findings identified meta managerial (MM) as the most impactful, meta contractual (MC) as the most uncertain, and meta financial (MF) and meta public authority (MP) as major barriers. Turkish respondents rated MM and MC 15%–28% lower than Americans, with larger gaps for MF and MP. Execution competency was assessed using Light Gradient Boosting Machine (LightGBM), with Q-factor and E-factor metrics used to test and validate the data. The final framework delivers a data-driven benchmark for guiding BIM integration in infrastructure-focused projects.