Fraud Detection using Risk Based Binary Classification

Abstract

A Google Pay fraud detection system is proposed in this research, implemented in C++, which detects fraudulent transactions through a risk-based binary classification framework. The system analyzes key factors such as transaction amount, transaction frequency, device trust, and location mismatch to determine transaction legitimacy A binary output is generated by the system, with 0 denoting legitimate transactions and 1 denoting fraudulent transactions. The study highlights how machine learning–inspired approaches can be applied to real-time payment systems to improve security and prevent financial losses.

Introduction

The text discusses the growing use of digital payment platforms such as Google Pay, emphasizing their role in improving transaction efficiency and convenience while also increasing exposure to fraud. Existing research highlights that cybercriminals exploit vulnerabilities in digital payment systems using advanced techniques, leading to unauthorized access, financial losses, and reduced user trust. As a result, the literature stresses the importance of robust security frameworks and effective fraud detection mechanisms.

It then explains the working of a Google Pay Fraud Detection System, which evaluates transaction legitimacy by analyzing key transactional and contextual features such as transaction amount, daily transaction frequency, device trust status, and location consistency. Each feature is weighted based on its fraud risk contribution, and a composite risk score is computed using a weighted sum. This score is converted into a fraud probability using a sigmoid function, and transactions are classified as fraudulent or legitimate based on a fixed threshold of 0.5. The system outputs interpretable risk scores and probabilities, supported by a sample dataset illustrating how inputs translate into fraud classification decisions.

Conclusion

The Google Pay detection system illustrates the effectiveness of risk-based classification in identifying fraudulent transactions. It delivers accurate results while requiring minimal resources and has the potential for further enhancement to support real-world deployment. This work underscores the critical role of fraud detection in digital payment systems and lays the groundwork for more advanced security solutions.

References

[1] Financial fraud detection through the application of machine learning techniques: a literature review [2] Artificial Intelligence in fraud detection: Revolutionizing financial security Prabin Adhikari , Prashamsa Hamal and Francis Baidoo Jnr , Lincoln University, California, USA. [3] https://ijsra.net/sites/default/files/IJSRA-2024-1860.pdf [4] https://datascience.codata.org/articles/10.5334/dsj-2023-023 [5] https://www.researchgate.net/publication/235701234_Fraud_Detection_and_Control_on_ATM_Machines_an_Algorithm_for_Combating_Cash_and_Fund_Transfer [6] https://www.irjet.net/archives/V9/i6/IRJET-V9I694.pdf [7] Fraud detection using machine learning by Aditya Oza chromextension://efaidnbmnnnibpcajpcglclefindmkaj/https://cs229.stanford.edu/proj2018/report/261.pdf [8] credit card fraud detection https://www.kaggle.com/datasets/mlg-ulb/creditcardfraud [9] Financial fraud detection applying data mining techniques: A comprehensive review from 2009 to 2019 , Volume 40, May 2021, 100402 [10] A Review of Data Mining-Based Financial Fraud Detection Research By Dianmin Yue; Xiaodan Wu; Yunfeng Wang; Yue Li; Chao-Hsien Chu [11] Credit Card Fraud Detection by kautalya sharma

Copyright

Copyright © 2026 Gurpreet Kaur. This 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.