Banking Intrusion Detection Systems based on customers behavior using Machine Learning algorithms: Comprehensive study

Abstract

In recent years, The computer networks has tremendous growth and the Internet usage became essential in many fields in real life, combined with the huge amount of data transmitted over networks, generated an exponential increase in the amount of malicious and ambiguous threats to computer networks. By implementation of Machine Learning (ML) algorithms to protect computer networks and to overcome network security breaches. Many approaches appeared to the surface to achieve that purpose, one of them is the Network Intrusion Detection System (NIDS). This research aims to present a comprehensive study about employing machine learning algorithms because of what they have of effective and productive characteristics and capabilities when used in the area of tracking user behavior embedded to construct a framework simulates a banking system and examines the deviation of customer's normal utilization of the system, that is called intrusion. The experimental results obtained by this research that involves combining four algorithms in one framework showed high accuracy and low false alarm detection rate.