DATA MINING AND TELECOMMUNICATION FRAUD DETECTION USING ARTIFICIAL NEURAL NETWORKS

摘要

This paper presents the use of neural networks for the detection of telecommunication fraud using the new concept of profiling. Fraud is a multi-billions dollar problem around the globe. The problem with telecommunication fraud is the huge loss of revenue and it can affect the credibility and performance of the telecommunication companies. The most difficult problem that faces the industry is the fact that fraud is dynamic. This means that whenever fraudesters feel that they will be detected, they find other ways to circumvent security measures. The advantage of using neural networks for detecting fraud is that the nonlinear nature and the parallel structure of the networks allow them to extract fraudulent behaviours without having to know the related equations. The network was successfully trained using simulated data resulting in a 98% detection rate when tested for call-sell fraud.