An Overview of ICT Frauds and their Detection with Bi-directional Artificial Neural Networks

Life and work of individuals and the overall society have become strongly dependant on information communication technology (ICT) systems and services. Due to their complexity, pretentiousness, need of convergence, demand to reduce the time-tomarket for new products and services, quality of standards and insufficient testing of end products, ICT systems are vulnerable to frauds. To prevent ICT frauds, it is necessary to identify and know them well. For this purpose we present the most usual and the most frequent ICT frauds and examples and methods used for their detection. Though every ICT fraud is specific, they altogether still have certain common properties, according to which we group them. In literature, ICT frauds are grouped into cloning fraud, toll fraud, subscriber fraud, social engineering fraud, computer intrusion fraud and credit card fraud. They are all discussed in this paper. Methods for detecting ICT frauds have been using artificial neural networks for quite some time. Instead of them we propose to employ sophisticated bi-directional artificial networks that were initially developed for other purposes. We introduce their basic working principle and their incorporation into the system for detecting ICT frauds. In the last section we discuss their adequacy for detecting individual ICT frauds.