Machine learning applications in risk management: Trends and research agenda

F1000Res. 2025 Feb 25;14:233. doi: 10.12688/f1000research.161993.1. eCollection 2025.

ABSTRACT

BACKGROUND: Currently, risk management is positioned as a key issue in industries, which is why machine learning technologies have been integrated for impact assessment, prevention, and decision making in different sectors. However, there are still important research gaps, so the aim is to investigate research trends related to the use of machine learning in risk management.

METHODS: A bibliometric analysis is proposed based on the PRISMA-2020 declaration in the Scopus and Web of Science databases.

RESULTS: The results show a growing interest in the use of machine learning for risk management in the scientific literature. China, South Korea and the United States lead the research. The thematic development reflects emerging topics such as urban trees and Covid-19. Key terms include random forest, SVM, and credit risk assessment, while others such as prediction, postpartum depression, big data, and security are considered emerging topics, reflecting the cross-cutting nature and applicability of the topic across different sectors of society. Deep learning and feature selection are also priorities for enhancing machine learning applications in risk management.

CONCLUSIONS: Machine learning in risk management has grown exponentially, shifting focus from stacking to urban trees and Covid-19. Key contributors, journals, and nations shape this evolving research landscape.

PMID:41059130 | PMC:PMC12498517 | DOI:10.12688/f1000research.161993.1