A Real-Time Prescriptive Solution for Explainable Cyber-Fraud Detection Within the iGaming Industry.

Area of Science:

• Cybersecurity
• Artificial Intelligence
• Data Science

Background:

• The iGaming industry faces significant challenges with cyber-fraud and player risk assessment.
• Traditional methods for fraud detection are often time-consuming and reactive.

Purpose of the Study:

• To develop a real-time, autonomous prescriptive solution for explainable cyber-fraud detection in iGaming.
• To leverage machine learning and explainable AI (XAI) for improved fraud prediction and assessment.

Main Methods:

• Utilized machine learning algorithms combined with eXplainable AI (XAI) techniques.
• Developed a prescriptive analytics pipeline for real-time player risk and fraud assessment.
• Implemented strategies to detect and manage concept drift in data.

Main Results:

• Achieved an average precision of 84.2% in predicting fraudulent behavior.
• Obtained an Area Under the Receiver Operating Characteristic curve (AUC ROC) of 0.82.
• Demonstrated the effectiveness of local interpretable explanations for proactive fraud fighting.

Conclusions:

• The developed solution offers a real-time, autonomous, and explainable approach to cyber-fraud detection in iGaming.
• Explainable AI provides actionable insights, enabling a proactive stance against fraud.
• The system effectively addresses concept drift, ensuring sustained performance.