The Role of Machine Learning in Risk Assessment and Management in Finance

Yousha, Saba and Khan, Ather Alam and Farooqi, Shaista Ashraf and Aslam, Rashid Ali and Shah, Syed Nadeem Juman (2025) The Role of Machine Learning in Risk Assessment and Management in Finance. Journal of Policy Research, 3 (2). pp. 680-686.

Text AIFinance.pdf Download (279kB)

Abstract

This paper analyzes the use of machine learning (ML) techniques for managing financial risks by applying supervised, unsupervised, and reinforcement learning to measure and control different types of financial risks like market, credit, and operational risk. The research design is blended, combining the quantitative analysis of the financial data sets with qualitative data collected from industry practitioners. The study gathered a diverse range of data from several sources, such as social media during foreign market hours, newsfeeds, financial market data, institutional loans, and other economic indicators to consider the impact of out-of-the-market factors. They were able to accurately predict asset prices and credit defaults utilizing machine learning methods like GBM and Neural Networks. At the same time, techniques such as K-Means Clustering and Autoencoders helped in finding out the market anomalies and concealed patterns. The Reinforcement Learning techniques, and in particular Deep Q-Networks (DQNs), were useful in modeling trading simulations and optimizing portfolio management strategies through real-time decision making. SHAP and LIME techniques were applied to model the predictions to make them lucid and enhance transparency and trust in the model. Although results were promising, there were concerns of data quality and extreme market conditions. Building real-time data integration, model strength during financial crisis, and data sharing through enhanced privacy measures should be the focus of future work. These findings are important for financial institutions because they bring so much insight in achieving Risk Management optimization through adopting more powerful less concealed and moreefficient financial risk management systems.

Item Type:	Journal
Uncontrolled Keywords:	Financial Risk Management, Machine Learning, Supervised Learning, Unsupervised Learning
Divisions:	School of Graduate Studies
Depositing User:	Muhamad Aizat Nazmi Mohd Nor Hamin
Date Deposited:	26 Aug 2026 01:59
Last Modified:	26 Aug 2026 01:59
URI:	http://ur.aeu.edu.my/id/eprint/1515

Actions (login required)

View Item