Machine learning applications in risk management: Trends and research agenda

Area of Science:

• * Computational Science and Engineering
• * Data Science and Analytics
• * Risk Management and Decision Science

Background:

• * Risk management is crucial across industries, increasingly leveraging machine learning (ML) for assessment and decision-making.
• * Existing literature has gaps in identifying emergent trends and cross-industry applications of ML in risk management.
• * Bibliometric analysis provides a systematic approach to map the evolving landscape of ML in risk management.

Purpose of the Study:

• * To conduct a bibliometric analysis of scientific literature on machine learning applied to risk management.
• * To identify key research trends, leading countries, and emerging application areas.
• * To map the evolution of methodologies from traditional to advanced techniques.

Main Methods:

• * Bibliometric analysis of scientific production sourced from Scopus and Web of Science databases.
• * Adherence to the PRISMA-2020 declaration for systematic literature review.
• * Identification and analysis of key terms, research trends, and country-specific contributions.

Main Results:

• * A significant surge in publications on ML for risk management, with a 98.99% growth between 2018 and 2023.
• * China, South Korea, and the United States identified as leading research contributors.
• * Emerging trends include ML for urban tree evaluation and SARS-CoV-2 risk management, with new focus areas like prediction, postpartum depression, big data, and security.

Conclusions:

• * The field of machine learning in risk management is rapidly expanding, with notable growth and international collaboration.
• * Methodologies are evolving, shifting from traditional approaches like stacking to advanced deep learning and feature selection.
• * Future research should explore novel applications and the integration of big data and security considerations.