FRAUD DETECTION IN MOBILE TELECOMMUNICATION
Fraud has been very common in our society, and it affects private enterprises as well as public entities. However, in recent years, the development of new technologies has also provided criminals more sophisticated way to commit fraud and it therefore requires more advanced techniques to detect and prevent such events. The types of fraud in Telecommunication industry includes: Subscription Fraud, Clip on Fraud, Call Forwarding, Cloning Fraud, Roaming Fraud, and Calling Card. Thus, detection and prevention of these frauds is one of the main objectives of the telecommunication industry. In this research, we developed a model that detects fraud in Telecommunication sector in which a random rough subspace based neural network ensemble method was employed in the development of the model to detect subscription fraud in mobile telecoms. This study therefore presents the development of patterns that illustrate the customers’ subscription's behaviour focusing on the identification of non-payment events. This information interrelated with other features produces the rules that lead to the predictions as earlier as possible to prevent the revenue loss for the company by deployment of the appropriate actions.