Year: 2025
Type: article
Abstract: As the global financial systems become more globalized and digitized, the threats of financial fraud are becoming massive accounting over $5 trillion annually, and sophisticated as cyber criminals ever-changing tactics. The traditional rule-based fraud detection systems that institutions have relied on are unable to meet this escalating challenge due to their inflexibility, high false-positive rates, and inability to process large-scale data in real-time. Machine learning offers an impressive alternative with transformative solutions that leverage predictive analytics models, anomaly detection, and adaptive learning to boost systems’ accuracy, scalability, and responsiveness in fraud detection. This article investigates the application of supervised, unsupervised, and deep learning machine learning techniques in detecting credit card fraud, identity theft, and cryptocurrency scams, highlighting their advantages over conventional methods. The article also highlights the challenges that machine learning techniques need to grapple with through more research and investment to be more efficient, such challenges include data imbalance, regulatory constraints, computational costs, and model interpretability.
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4409478483