Predicting online banking fraud using Adaptive Neuro-Fuzzy Inference System (ANFIS)

Abstract

Online banking is growing tremendously in recent years. This is attributed to the explosion in internet technologies on both computer and handheld-devices platforms. Unfortunately, the growth of online banking is accompanied by a persistent growth of attacks in the form of phishing and malware. Therefore, ensuring security of personal data during transactions is of utmost concern. Security standards have become more stringent in order to counter the growth rate of online fraud. Multiple defensive security lines have also been proposed to reduce the rate of attacks. Customer and vendor awareness are considered important in the fight against online fraud. Technical solutions such as online bank coders/decoders are further presented as a solution to fraud. Nevertheless, despite all the efforts involved, fraud is still increasing. This relentless growth in online fraud has provoked this work, which seeks to deal with fraud from a different perspective. This work undertakes the task of forecasting fraud volume in the future based on the rate of growth of attacks in the last decade, in relation to technological and social developments. Technological developments are characterized by the systems used in accessing the Internet and their security flaws, while social development is represented by population growth and the increasing reliance on internet facilities. In this research a system has been developed to help achieve a clear picture of the volume of fraud and the factors affecting its growth in the future. This will also help developers to have more time to prepare their defensive tools, as well as help the management of banking institutions to establish solid plans to fight against fraud. The major findings of this study indicate that personal computer operating system usage rate, Mobile operating system (such as IOS and Android) and web browser types are the main risk factors of online banking fraud. The findings also reflect that online banking fraud growth is related to social distribution and technology advancement. Finally, there is a need for awareness of data sufficiency to build a prediction system which can forecast the volume of fraud expected in the future