Big data analytics framework for real-time fraud detection in public financial systems

Fraud in public financial systems has become a critical challenge due to the rapid expansion of digital government services and the increasing volume of high-frequency financial transactions. Traditional fraud detection mechanisms, which rely on rules and batch processing, are no longer sufficient to address the complexity, speed, and evolving nature of modern fraudulent activities. This study proposes a big data analytics framework for real-time fraud detection in public financial systems by integrating streaming analytics, machine learning algorithms, and distributed computing architectures. The framework enables continuous monitoring of financial transactions, allowing for immediate detection of anomalies and fraudulent behavior with minimal latency. It incorporates supervised learning models for classification, unsupervised anomaly detection techniques for unknown fraud patterns, and hybrid ensemble approaches to improve detection robustness. Additionally, streaming data processing ensures scalability and real-time responsiveness in large-scale government financial infrastructures. The proposed model is expected to enhance detection accuracy, reduce false positives, and strengthen the overall security and transparency of public financial systems. The study contributes to advancing intelligent financial security systems and supports the development of adaptive, scalable, and real-time fraud prevention mechanisms for modern digital governance environments.