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

Area of Science:

• Occupational Health
• Industrial Safety
• Machine Learning Applications

Background:

• Sleep deprivation significantly impacts construction worker health, safety, and job performance.
• Existing research on sleep deprivation primarily focuses on cognitive impairment's effect on safety and productivity.
• Limited studies explore the direct link between sleep deprivation, workplace hazards, and injury causation.

Purpose of the Study:

• To address the gap in understanding sleep deprivation's role in construction workplace hazards.
• To utilize machine learning algorithms for predicting hazardous situations stemming from sleep deprivation.
• To identify key predictive factors associated with sleep deprivation in construction workers.

Main Methods:

• Employed machine learning algorithms, specifically Support Vector Machine (SVM) and Random Forest.
• Developed a predictive model for sleep deprivation using data from 240 construction workers.
• Identified seven primary indices as significant predictive factors for sleep deprivation.

Main Results:

• The Support Vector Machine (SVM) algorithm demonstrated superior performance in predicting sleep deprivation during validation.
• Seven key factors were identified as significant predictors of sleep deprivation in the construction workforce.
• The study successfully demonstrated the applicability of machine learning for predicting sleep deprivation.

Conclusions:

• Machine learning offers a viable tool for the timely and accurate prediction of sleep deprivation in construction.
• These predictive insights can empower stakeholders, like safety managers, to implement targeted interventions.
• Proactive identification and management of sleep deprivation can lead to reduced accidents and enhanced construction site safety.