Evaluating Machine Learning Based Behavioral Analytics for Detecting Unusual Transactional Sequences in High-Velocity UPI Digital Payment Networks

  • Unique Paper ID: 211387
  • PageNo: 43-49
  • Abstract:
  • The Unified Payments Interface (UPI) allows fast bank-to-bank payments. This speed makes early fraud detection important. A payment may look normal on its own but become suspicious when compared with a user’s recent activity. This paper proposes a simple way to study such patterns using machine learning. The system will use transaction details and summaries of earlier payments, such as the number of recent transfers, changes in spending, new payees and the time between payments. A rule-based method, Logistic Regression, Random Forest and Isolation Forest will be compared. Only one selected model will be used in the main scoring path. The proposed architecture has four steps: collect data, prepare behaviour features, calculate a risk score and flag suspicious payments for review. Testing will keep transactions in time order and measure fraud detection, false alerts and processing time. Public mobile-money data and clearly labelled synthetic UPI-like data will be used where suitable. Results from these sources will not be presented as proof of performance on live UPI payments. The study aims to find whether simple behaviour features can improve fraud detection without a complex system.

Copyright & License

Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

BibTeX

@article{211387,
        author = {Mrs. Aarti Atul Pathade},
        title = {Evaluating Machine Learning Based Behavioral Analytics for Detecting Unusual Transactional Sequences in High-Velocity UPI Digital Payment Networks},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {no},
        pages = {43-49},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=211387},
        abstract = {The Unified Payments Interface (UPI) allows fast bank-to-bank payments. This speed makes early fraud detection important. A payment may look normal on its own but become suspicious when compared with a user’s recent activity. This paper proposes a simple way to study such patterns using machine learning. The system will use transaction details and summaries of earlier payments, such as the number of recent transfers, changes in spending, new payees and the time between payments. A rule-based method, Logistic Regression, Random Forest and Isolation Forest will be compared. Only one selected model will be used in the main scoring path. The proposed architecture has four steps: collect data, prepare behaviour features, calculate a risk score and flag suspicious payments for review. Testing will keep transactions in time order and measure fraud detection, false alerts and processing time. Public mobile-money data and clearly labelled synthetic UPI-like data will be used where suitable. Results from these sources will not be presented as proof of performance on live UPI payments. The study aims to find whether simple behaviour features can improve fraud detection without a complex system.},
        keywords = {UPI; fraud detection; machine learning; payment behaviour; unusual transaction patterns; false alerts; processing time.},
        month = {November},
        }

Cite This Article

Pathade, M. A. A. (2026). Evaluating Machine Learning Based Behavioral Analytics for Detecting Unusual Transactional Sequences in High-Velocity UPI Digital Payment Networks. International Journal of Innovative Research in Technology (IJIRT), 43–49.

Related Articles