Machine learning in Occupational Safety and Health - a systematic review

• Type: Article in International Scientific Journal
• Date: 2023

Authors

Maheronnaghsh, S (Author) Other; Zolfagharnasab, H (Author) Other; Gorgich, M (Author) Other; Duarte, J (Author) FEUP

Journal

Title: International Journal of Occupational and Environmental Safety

Vol. 21

Pages: 14-32

ISSN: 2184-0954

Publisher: FEUP - Sociedade Portuguesa de Segurança e Higiene Ocupacionais

Other information

Authenticus ID: P-016-17K

DOI: 10.24840/2184-0954_007-001_001586

Abstract (EN): <jats:p>With the development of technology, machine learning (ML), a branch of computer science that aims to turn computers into decision-making agents using the most appropriate algorithms, is also paving its way in the modern world. This systematic review arises from the need to understand the impact and report the best practices for applying ML in occupational safety and health. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines were used to provide the best research results. From the 759 identified papers, only 42 were included in the study after applying both exclusion and inclusion criteria. Application is primarily used in accident and risk assessment, and construction and office work are the leaders in applications. The applied methods mainly consist of classification (injuries, accidents, monitoring data), prediction (of hazards), and regression (to find patterns of accidents to prevent them). In conclusion, decision-makers and workers are taking advantage of various artificial intelligence techniques to find solutions in the occupational safety and health environment when experts have access to correct data, either in real-time or recorded datasets. However, it is necessary that in future investigations, limitations of using ML applications in occupational safety and health area be improved and their full potential is achieved.</jats:p>

Language: English

Type (Professor's evaluation): Scientific