I. INTRODUCTION
Traditionally, banks do not publicise specific details of their Fraud Detection (FD) systems [1]. Despite this fact, there have recently been a number of software vendors who publicised some information about their Internet banking FD tools and the market share of such tools in commercial banks worldwide. For example, the Proactive Risk Manager (PRM) was reported to be used by the top 20 banks in the world and in more than 40 countries [2]. Similarly, the SAS Fraud Management system is reportedly used in debit and credit card FD solutions at more than 43000 online banking sites [3]. Although the information publicised is relatively simplified and primarily released for marketing purposes, the details exposed give a good indication of the generic architecture used by commercial banks to detect fraud in online banking transactions. This research has found out from different software vendors' white papers that there is a consistent use of a combination of a Rule-Based System (RBS) with an Artificial Neural Network (ANN). The Falcon Fraud Manager, a payment card fraud detection tool comprises an RBS and ANN as its main components [4]. The PRM white paper likewise reports that the PRM debit/credit card and internet banking system uses a hybrid structure featuring an RBS and ANN [2]. In a similar fashion, the SAS Fraud Management system's architecture is reported to mainly include an RBS and an ensemble of Self Organising Neural Networks (SONNA) [3].