AI-Enhanced Data Engineering for Real-Time Fraud Detection in Digital Transactions

Abstract

In the evolving landscape of digital transactions, the proliferation of online financial activities has 
heightened the risk of fraud, necessitating advanced methods for real-time fraud detection. This 
paper explores the application of AI-enhanced data engineering techniques to improve the 
detection and prevention of fraudulent activities in digital transactions. We present a 
comprehensive framework integrating machine learning algorithms, data preprocessing 
techniques, and real-time analytics to identify and mitigate fraudulent behavior. Utilizing a dataset 
comprising transaction records from various financial institutions, we implement several AI 
models, including supervised learning algorithms and anomaly detection techniques, to evaluate 
their effectiveness in fraud detection. Our results demonstrate significant improvements in 
detection accuracy, with the AI-enhanced system achieving a reduction in false positives and a 
higher true positive rate compared to traditional methods. The study highlights the potential of 
combining AI with robust data engineering practices to enhance security measures in digital 
transactions, providing a foundation for future advancements in fraud detection technologies.