Construction accident prevention: A systematic review of machine learning approaches

Area of Science:

• Construction Safety
• Machine Learning Applications
• Accident Prevention

Background:

• The construction industry faces high work-related accident rates, necessitating improved safety management.
• Technological advancements, including machine learning (ML), offer potential for optimizing construction processes and decision-making.
• Advanced research is crucial to understand how ML can best be applied to enhance construction site safety.

Purpose of the Study:

• To conduct a systematic literature review on machine learning (ML) applications in construction accident prevention.
• To identify and address knowledge gaps in the current research landscape.
• To retrieve relevant publications from ten scientific databases.

Main Methods:

• Bibliometric research and descriptive analysis of 73 scientific articles.
• Systematic literature review across ten scientific databases.
• Examination of publication trends, key contributors, and research methodologies.

Main Results:

• The USA and China lead in publications related to ML in construction safety.
• Support Vector Machine (SVM) is the most frequently employed ML method.
• Textual data, including inspection reports and accident narratives, are commonly used, focusing on proactive, pre-accident data.

Conclusions:

• The review provides insights into current ML applications for construction safety management.
• Identified areas for improvement and future research directions in the field.
• Highlights the potential of ML in enhancing safety protocols and reducing accidents in the construction industry.