Application of Machine Learning Methods to Detect Fraud in Bank Transactions
DOI:
https://doi.org/10.52575/2687-0932-2026-53-1-111-121Keywords:
fraudulent transactions, intelligent systems, machine learning methods, model accuracy, algorithm efficiency, cross-validationAbstract
The article touches on the connection between the mass distribution of intellectual technologies and their possible increased use in illegal actions. The authors focus on development and analysis of machine learning methods for detecting fraudulent transactions in the financial sector. The relevance of the topic is explained by the ongoing evolution of fraudulent schemes, which requires innovative technologies to be applied for ensuring financial security. The paper considers modern approaches to data processing, scaling and elimination of data asymmetry. The study covers model training using four algorithms: logistic regression, decision tree, random forest method, and gradient descent. To assess the quality of the model, the ROC-AUC metric was used, as well as characteristics such as accuracy, completeness and F1 measure. The logistic regression model performed best, achieving a ROC-AUC value of 0.975 on the test dataset. The results of the work highlight the practical value of machine learning models as a reliable tool for minimizing the risks associated with fraudulent transactions.
Downloads
References
Список литературы
Аскаров Е.Ф., Хамитов Р.М. 2024. Использование временных рядов для прогнозирования мошеннических операций. Экономика и предпринимательство, 3(164): 1356–1359.
Григорьев А. 2023. Машинное обучение. Портфолио реальных проектов. Санкт-Петербург: Питер, 496 с.
Мартин Р. 2022. Чистая архитектура. Искусство разработки программного обеспечения. Санкт-Петербург: Питер, 352 с.
Марченко А.Л. 2023. Python: большая книга примеров. Издательство Московского университета, 361 с.
Окуньков С.В., Барулина М.А., Санбаев А.К. 2023. Мультиклассовая классификация на сильно несбалансированном наборе данных. Фундаментальная и прикладная медицина: материалы Международной конференции молодых ученых, Саратов, 105–106.
Орельен Ж. 2020. Прикладное машинное обучение с помощью Scikit-Learn, Keras и TensorFlow: концепции, инструменты и методы построения интеллектуальных систем, 2-е изд. Санкт-Петербург: ООО "Диалектика", 1520 с.
Плас Дж. В. Python для решения сложных задач: наука о данных и машинное обучение. Санкт-Петербург: Питер, 2021. 576 с.
Траск Э. 2025. Грокаем глубокое обучение. Санкт-Петербург: Питер, 352 с.
Хлобыстова А.О., Абрамов М.В. 2024. Публичность организации как уязвимость при проведении социоинженерной атаки. Информационное общество, 1: 85–93.
Chio K. 2020. Machine learning and security. Protecting systems with data and algorithms. Moscow: DMK Press, 388 p.
ICO Falcon Fraud Manager [Electronic resource]. URL: https://www.fico.com/en/products/fico-falcon-fraud-manager (date of request: 20.10.2025)
Ioffe L. 2024. Application of big data technology for fraud detection in financial transactions. Universum: technical sciences, 2(119): 6.
Kelleher J.D. 2019. Deep Learning. The Massachusetts Institute of Technology, 296 p.
Liu Yu., Li Ya., Xie D. 2024. Implications of imbalanced datasets for empirical ROC-AUC estimation in binary classification tasks. Journal of Statistical Computation and Simulation, 94(1): 183–203.
Madani A. 2023. Debugging Machine Learning Models with Python. Develop high-performance, low bias, and explainable machine learning and deep learning models. Birmingham: Packt Publishing Ltd, 344 p.
Omolara O., Agwubuo C., Onyeche S., Omotoyosi O., Kenneth N. and Olajumoke A. 2024. The impact of big data analytics on financial risk management. International Journal of Science and Research Archive, 12(02): 821–827.
Wang Y., Wang Q., Zhao L., Wang C. 2023. Differential privacy in deep learning: Privacy and beyond. Future Generation Computer Systems, 148: 408–424.
Ye J.X. 2023. A review of two-stage target detection algorithms based on deep learning. Internet Wkly, 18: 16–18.
Yuxi (Hayden) Liu. 2020. Python Machine Learning By Example. Third Edition. Build intelligent systems using Python, Tensor Flow 2, PyTorch, and scikit-learn. Birmingham: Packt Publishing Ltd, 526 p.
Zhu H., Zhou S.Y. 2023. A review of single-stage target detection algorithms based on deep learning. Ind. Control. Comput. 36: 101–103.
References
Askarov E.F., Xamitov R.M. 2024. Ispol`zovanie vremenny`x ryadov dlya prognozirovaniya moshennicheskix operacij [Using time series to predict fraudulent transactions]. E`konomika i predprinimatel`stvo, 3(164):1356–1359.
Grigor`ev A. 2023. Mashinnoe obuchenie. Portfolio real`ny`x proektov [Machine Learning. Portfolio of Real Projects]. Sankt-Peterburg: Piter, 496 s.
Martin R. 2022. Chistaya arxitektura. Iskusstvo razrabotki programmnogo obespecheniya [Clean Architecture. The Art of Software Development]. Sankt-Peterburg: Piter, 352 s.
Marchenko A.L. 2023. Python: bol`shaya kniga primerov [Python: a great book of examples]. Izdatel`stvo Moskovskogo universiteta, 361 s.
Okun`kov S.V., Barulina M.A., Sanbaev A.K. 2023. Mul`tiklassovaya klassifikaciya na sil`no nesbalansirovannom nabore danny`x [Multiclass classification on a highly imbalanced dataset]. Fundamental`naya i prikladnaya medicina: materialy` Mezhdunarodnoj konferencii molody`x ucheny`x, Saratov, 105–106.
Orel`en Zh. 2020. Prikladnoe mashinnoe obuchenie s pomoshh`yu Scikit-Learn, Keras i TensorFlow: koncepcii, instrumenty` i metody` postroeniya intellektual`ny`x sistem [Applied Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Methods for Building Intelligent Systems], 2-e izd. Sankt-Peterburg: OOO Dialektika, 1520 s.
Plas Dzh. V. Python dlya resheniya slozhny`x zadach: nauka o danny`x i mashinnoe obuchenie [Python for solving complex problems: data science and machine learning]. Sankt-Peterburg: Piter, 2021. 576 s.
Trask E`. 2025. Grokaem glubokoe obuchenie [Grok deep learning]. Sankt-Peterburg: Piter, 352 s.
Xloby`stova A.O., Abramov M.V. 2024. Publichnost` organizacii kak uyazvimost` pri provedenii socioinzhenernoj ataki [The organization's public nature as a vulnerability in a social engineering attack]. Informacionnoe obshhestvo, 1:85–93.
Chio K. 2020. Machine learning and security. Protecting systems with data and algorithms. Moscow: DMK Press, 388 p.
ICO Falcon Fraud Manager [Electronic resource]. URL: https://www.fico.com/en/products/fico-falcon-fraud-manager (date of request: 20.10.2025)
Ioffe L. 2024. Application of big data technology for fraud detection in financial transactions. Universum: technical sciences, 2(119): 6.
Kelleher J.D. 2019. Deep Learning. The Massachusetts Institute of Technology, 296 p.
Liu Yu., Li Ya., Xie D. 2024. Implications of imbalanced datasets for empirical ROC-AUC estimation in binary classification tasks. Journal of Statistical Computation and Simulation, 94(1): 183–203.
Madani A. 2023. Debugging Machine Learning Models with Python. Develop high-performance, low bias, and explainable machine learning and deep learning models. Birmingham: Packt Publishing Ltd, 344 p.
Omolara O., Agwubuo C., Onyeche S., Omotoyosi O., Kenneth N. and Olajumoke A. 2024. The impact of big data analytics on financial risk management. International Journal of Science and Research Archive, 12(02): 821–827.
Wang Y., Wang Q., Zhao L., Wang C. 2023. Differential privacy in deep learning: Privacy and beyond. Future Generation Computer Systems, 148: 408–424.
Ye J.X. 2023. A review of two-stage target detection algorithms based on deep learning. Internet Wkly, 18: 16–18.
Yuxi (Hayden) Liu. 2020. Python Machine Learning By Example. Third Edition. Build intelligent systems using Python, Tensor Flow 2, PyTorch, and scikit-learn. Birmingham: Packt Publishing Ltd, 526 p.
Zhu H., Zhou S.Y. 2023. A review of single-stage target detection algorithms based on deep learning. Ind. Control. Comput. 36: 101–103.
Abstract views: 6
Share
Published
How to Cite
Issue
Section
Copyright (c) 2026 Economics. Information Technologies
This work is licensed under a Creative Commons Attribution 4.0 International License.
