A2 Refereed review article in a scientific journal
Machine learning and deep learning for safety applications: Investigating the intellectual structure and the temporal evolution
Authors: Leoni Leonardo, Bahootoroody Ahmad, Abaei Mohammad Mahdi, Cantini Alessandra, Bahootoroody Farshad, De Carlo Filippo
Publisher: ELSEVIER
Publishing place: AMSTERDAM
Publication year: 2024
Journal: Safety Science
Journal name in source: SAFETY SCIENCE
Journal acronym: SAFETY SCI
Article number: 106363
Volume: 170
Number of pages: 25
ISSN: 0925-7535
eISSN: 1879-1042
DOI: https://doi.org/10.1016/j.ssci.2023.106363
Publication's open availability at the time of reporting: Open Access
Publication channel's open availability : Partially Open Access publication channel
Web address : https://doi.org/10.1016/j.ssci.2023.106363
Self-archived copy’s web address: https://research.utu.fi/converis/portal/detail/Publication/387113758
Over the last decades, safety requirements have become of primary concern. In the context of safety, several strategies could be pursued in many engineering fields. Moreover, many techniques have been proposed to deal with safety, risk, and reliability matters, such as Machine Learning (ML) and Deep Learning (DL). ML and DL are characterised by a high variety of algorithms, adaptable for different purposes. This generated wide and fragmented literature on ML and DL for safety purposes, moreover, literature review and bibliometric studies of the past years mainly focus on a single research area or application field. Thus, this paper aims to provide a holistic understanding of the research on this topic through a Systematic Bibliometric Analysis (SBA), along with proposing a viable option to conduct SBAs. The focus is on investigating the main research areas, application fields, relevant authors and studies, and temporal evolution. It emerged that rotating equipment, structural health monitoring, batteries, aeroengines, and turbines are popular fields. Moreover, the results depicted an increase in popularity of DL, along with new approaches such as deep reinforcement learning through the past four years. The proposed workflow for SBA has the potential to benefit researchers from multiple disciplines, beyond safety science.
Downloadable publication This is an electronic reprint of the original article. |  |
