Enhancing construction safety: predicting worker sleep deprivation using machine learning algorithms

Abstract

Sleep deprivation is a critical issue that affects workers in numerous industries, including construction. It adversely affects workers and can lead to significant concerns regarding their health, safety, and overall job performance. Several studies have investigated the effects of sleep deprivation on safety and productivity. Although the impact of sleep deprivation on safety and productivity through cognitive impairment has been investigated, research on the association of sleep deprivation and contributing factors that lead to workplace hazards and injuries remains limited. To fill this gap in the literature, this study utilized machine learning algorithms to predict hazardous situations. Furthermore, this study demonstrates the applicability of machine learning algorithms, including support vector machine and random forest, by predicting sleep deprivation in construction workers based on responses from 240 construction workers, identifying seven primary indices as predictive factors. The findings indicate that the support vector machine algorithm produced superior sleep deprivation prediction outcomes during the validation process. The study findings offer significant benefits to stakeholders in the construction industry, particularly project and safety managers. By enabling the implementation of targeted interventions, these insights can help reduce accidents and improve workplace safety through the timely and accurate prediction of sleep deprivation.