Real-Time Fraud Prediction On Streaming Customer-Behaviour Data

The field of detection and analysis of fraud situations continues to be an essential topic in the telecom industry. This study proposes a methodology that can predict abnormal behavior situations on real-time streaming customer behavior data in the telecommunication domain. The prototype implementation of the proposed method is designed, developed, and applied to the telecommunications dataset. We perform performance tests for the method's prediction success and scalability metrics on the designed prototype application. The results indicate that the proposed method proves to be a promising approach in the telecommunication sector.