2025, Vol. 12, No. s3. - go to content...

Permanent address of this page - https://resources.today/en/03faor325.html

Метаданные этой статьи так же доступны на русском языке

DOI: 10.15862/03FAOR325 (https://doi.org/10.15862/03FAOR325)

Full article in PDF format (file size: 483.5 KB)

For citation:

Bestaev G.B. Application of machine learning methods to predict cyber threats in financial ecosystems. Russian journal of resources, conservation and recycling. 2025; 12(s3). Available at: https://resources.today/PDF/03FAOR325.pdf (in Russian). DOI: 10.15862/03FAOR325

Application of machine learning methods to predict cyber threats in financial ecosystems

Bestaev Georgy Badrievich
Financial University under the Government of the Russian Federation, Moscow, Russia
E-mail: 224944@edu.fa.ru

Abstract. The growing complexity of the cyber threat landscape and the intensification of targeted attacks on the financial sector necessitate a transition from a reactive to a proactive information security model based on predicting probable incident scenarios. Machine learning methods provide financial ecosystems with tools for analyzing transactional, network, and behavioral data sets, identifying nontrivial patterns, and generating cyber risk probability models. The purpose of this study is to identify the specifics of applying machine learning methods to predict cyber threats in financial ecosystems, explore their capabilities, and assess the key limitations of their practical implementation. This paper systematizes the main approaches to cyberthreat forecasting using machine learning. It also conducts a comparative analysis of algorithm classes (supervised learning, unsupervised learning, and deep learning) based on their applicability to financial data. It also proposes a classification of limitations to the implementation of intelligent models, differentiated by the nature of the barrier. The results demonstrate that the greatest effectiveness is achieved through the integrated use of several classes of models integrated into the overall cyber risk management architecture, provided that the organization has a combination of technological maturity, sufficient data completeness, and the existence of regulations defining the role of intelligent solutions in the decision-making system. The scientific novelty lies in the development of a classification of limitations to the implementation of ML models in the financial sector, differentiated by the nature of the barrier, and in the systematization of forecasting approaches, taking into account the specifics of the Russian cyber threat landscape for 2025–2026. The practical significance lies in the potential for financial institutions to use the findings when formulating strategies for implementing intelligent threat detection systems and developing regulations governing the role of ML models in decision-making.

Keywords: machine learning; cyberthreats; financial ecosystems; cybersecurity; forecasting; supervised learning; unsupervised learning; deep learning; antifraud; data mining

Download article in PDF format