Student Research Spotlight: Using Machine Learning to Prevent Credit Card Fraud

A Data-Driven Look at Fraud Prevention

Onat and Dr. Back studied more than 568,000 anonymized credit card transactions, applying machine learning models to detect patterns associated with fraud. Their goal was to understand how different types of cyber guardianship — such as bank-level fraud monitoring, multi-factor authentication, and user security behaviors — influence the likelihood of fraudulent activity.

“Dr. Back and I were both really interested in how technology and crime overlap, especially with how common online banking and credit card use have become, and especially with banks being a huge target for cyber criminals,” Onat says.

The Role of Cyber Guardianship in Modern Cybersecurity

The study focused on the concept of cyber guardianship, a digital extension of crime prevention theory. Guardianship includes:

• Secure passwords and authentication steps
• Real-time account monitoring
• Machine learning systems that flag anomalies
• User awareness and safe online habits
• AI-driven fraud detection tools

“We wanted to see if the same ideas that explain how people protect themselves in the real world could also apply online,” Onat says. “That’s where the idea of cyber guardianship came in — anything that helps protect people digitally, like fraud alerts, secure passwords, or AI systems that detect suspicious activity.”

Boosting Banking Security with AI-Driven Fraud Detection

Their results confirmed that strong digital guardianship significantly reduces fraud, supporting the importance of cybersecurity best practices at both the institutional and individual level.

“Stronger digital ‘guards’ really do protect both banks and customers,” Onat says. “In this instance, the ‘guard’ would be implementing machine learning systems.”

“Banks already use machine learning to flag unusual spending or alert customers if something looks off,” she says. “Our research helped show that these tools, when used properly, really do make a difference in lowering the risk of fraud.”

Graduate-Level Cybersecurity Research Makes an Impact

Onat’s experience demonstrates the power of faculty mentorship and research opportunities.

“Working with Dr. Back has been one of the best parts of my time at Scranton. It gave me real research experience and showed me how classroom knowledge applies to real problems…That perspective has helped me think more critically and practically as a grad student,” Onat says.

As emerging threats continue to reshape the cybersecurity landscape, opportunities like Onat’s show how graduate-level research can prepare students to meet those challenges with confidence.

In The University of Scranton’s MS in Cybercrime Investigation and Cybersecurity program, students gain direct experience with cutting-edge tools while working side-by-side with expert faculty. The program equips future professionals with the skills, insight, and hands-on experience needed to make an immediate impact in the field.

Ready to advance your career? Explore the Cybercrime Investigation and Cybersecurity master's today!