Unraveling safety risk coupling in prefabricated steel construction via data mining

As an important form of industrialized construction, prefabricated steel buildings have promoted the sustainable development of the construction industry. However, the construction safety issues of prefabricated steel buildings have not received enough attention. The safety risk characteristics are still unclear, and the complex interrelationships between various risk factors have not been fully explored. To address the limitations of existing research, this paper employs the BM25 algorithm to identify risk factors, followed by the frequent pattern (FP)-growth algorithm to analyze associations among these factors. Finally, complex network analysis is used to reveal the coupling mechanisms of safety risks in prefabricated steel construction. The study finds that the hoisting phase is a peak period for safety accidents, with high-altitude operations being the key scene in which safety risks are most concentrated. Additionally, the safety risks of prefabricated steel construction are not triggered by a single factor but rather exhibit a significant multidimensional coupling characteristic involving human, material/equipment, site, and management. Among these, the inadequate on-site supervision, improper operation, and inadequate safety education and training are the key coupling factors that lead to safety accidents. This study contributes by proposing novel text mining methods and association rule mining techniques, providing new perspectives and approaches for the identification and analysis of safety risks in prefabricated steel construction. Based on the research results, this study proposes specific safety prevention and control measures, providing targeted guidance for project implementation.