Adaptive UPI Fraud Detection: A Hydrid Machine Learning Framework

DOI

10.2991/978-94-6463-858-5_51How to use a DOI?

Keywords

UPI Fraud; Machine Learning; Hidden Markov Model; Convolutional Neural Networks; Real-time Detection; Anomaly Detection; Financial Security; Data Imbalance; Fraud Prevention Introduction

Abstract

The increasing reliance on Unified Payments Interface (UPI) for digital transactions has significantly transformed the financial landscape, making digital payments seamless and accessible. However, this widespread adoption has also led to an escalation in fraudulent activities, including phishing, social engineering, fake UPI handles, SIM cloning, and malware-based attacks, causing notable financial damages. To tackle this issue, this research proposes a hybrid machine learning approach that integrates Convolutional Neural Networks (CNNs) and Hidden Markov Models (HMMs) for efficient fraud detection. The HMM is employed to study transactional sequences, detecting unusual deviations from a user’s established patterns, whereas CNNs analyze transaction characteristics to identify intricate fraudulent patterns, thereby strengthening anomaly detection. One of the primary obstacles in fraud detection is the inherent imbalance in transaction datasets, where fraudulent cases constitute a significantly smaller proportion compared to legitimate ones. Additionally, the continuous evolution of fraud strategies and the challenge of false positives further complicate detection. To address these challenges, the study incorporates multiple machine learning techniques, advanced resampling methods, and refined feature engineering to enhance prediction accuracy. By leveraging real-time analytics and publicly available datasets, this approach establishes a scalable and adaptive fraud detection model, enhancing security, minimizing financial threats, and fostering trust within the digital payment ecosystem.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Tholapi Sri Kartik
AU  - G. Charan Teja
AU  - Williams
AU  - K. Raghu
PY  - 2025
DA  - 2025/11/04
TI  - Adaptive UPI Fraud Detection: A Hydrid Machine Learning Framework
BT  - Proceedings of International Conference on Computer Science and Communication Engineering (ICCSCE 2025)
PB  - Atlantis Press
SP  - 591
EP  - 602
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6463-858-5_51
DO  - 10.2991/978-94-6463-858-5_51
ID  - Kartik2025
ER  -