Malik Ahsan Arif

Developing-country construction sectors continue to record disproportionately high occupational accident rates, partly attributable to the slow adoption of digital safety technologies, including Building Information Modeling (BIM) and Internet of Things (IoT) systems. While prior empirical research has established the population-level factors that explain stakeholder adoption intention through survey-based frameworks, the ability to classify individual stakeholder readiness for targeted, pre-deployment intervention remains methodologically unaddressed. This study fills that gap by applying three supervised machine learning classifiers (Random Forest [RF], XGBoost (XGB), and Support Vector Machine (SVM)) to a dataset of 107 construction professionals purposively sampled from large-scale infrastructure projects in Pakistan, including China−Pakistan Economic Corridor (CPEC) packages and the Barakahu Bypass project. Five construct-level features derived from an integrated Technology Acceptance Model and Technology−Organization−Environment (TAM-TOE) survey instrument were used to classify stakeholders into High, Moderate, and Low readiness tiers. XGBoost achieved the best classification performance (accuracy = 93%, macro F1 = 0.93), followed by RF (91%, F1 = 0.91) and SVM (87%, F1 = 0.87). The convergent performance across three structurally different algorithm families indicates that the readiness signal reflects a consistent attitudinal pattern rather than an artifact of any single modeling assumption. Feature importance analysis consistently identified Perceived Benefits (32%) and Technology Awareness (25%) as the dominant predictive features, followed by Organizational Readiness (20%), Perceived Barriers (15%), and Respondent Profile (8%). Attitudinal readiness mapping classified 62% of stakeholders as High readiness, 28% as Moderate, and 10% as Low, providing an exploratory attitudinal segmentation framework to assist construction managers in prioritizing capacity-building investments, subject to longitudinal behavioral validation. The study also finds that awareness of digital technology consistently outpaces Organizational Readiness for implementation, a pattern consistent with findings from analogous developing-country construction contexts.

Construction sectors in developing countries continue to experience disproportionately high fatality rates, largely due to reactive safety practices and the limited adoption of digital safety technologies. While Building Information Modeling (BIM) and the Internet of Things (IoT) offer significant potential for proactive safety management, their integrated application remains underexplored in resource-constrained contexts. This study examines stakeholder readiness to adopt IoT-enhanced BIM-based safety monitoring systems in large-scale infrastructure projects in Pakistan, including China–Pakistan Economic Corridor (CPEC) initiatives and the Barakahu Bypass project. An integrated Technology Acceptance Model–Technology–Organization–Environment (TAM–TOE) readiness framework is employed, wherein TAM-derived cognitive-motivational factors (Technology Awareness and Perceived Benefits) and TOE-derived contextual factors (Organizational Readiness and Perceived Barriers) are examined as joint predictors of Behavioral Intention (BI). Data were collected from 107 purposively sampled construction professionals using a structured questionnaire. The results indicate high attitudinal readiness (BI mean = 4.7; perceived benefits mean = 4.6) alongside moderate organizational readiness (mean = 3.4). Regression analysis reveals that perceived benefits (β = 0.42, p < 0.001) and technology awareness (β = 0.29, p = 0.003) are the strongest positive predictors of adoption intention. In contrast, perceived barriers exert a significant negative effect (β = −0.22, p = 0.022). The model explains 61.2% of the variance in behavioral intention. This study advances the literature by providing empirical evidence on stakeholder readiness for BIM–IoT safety adoption within construction management processes, estimated through a multiple regression model. It offers practical implications for policymakers and industry stakeholders seeking to accelerate data-driven decision-making and digital safety transformation in developing economies.