Soft computing methods in cybersecurity

Abstract

In this paper, the aim is to show how computational intelligence methods (fuzzy logic, neural networks, genetic algorithms) can be used for detecting and preventing the most common attacks (Evasion, Data Poisoning, DoS and DDoS) to test its performance and weaknesses against adopted malware and to identify advanced persistent threats early on its life cycle using prescriptive analytics. This research also propose some soft computing based general solutions to minimize the alert fatigue caused by generating too many false alerts by getting insights from security information and event management (SIEM) signals to make the prevention systems learn from the environment to keep false signals to a minimum

In this paper, the aim is to show how computational intelligence methods (fuzzy logic, neural networks, genetic algorithms) can be used for detecting and preventing the most common attacks (Evasion, Data Poisoning, DoS and DDoS) to test its performance and weaknesses against adopted malware and to identify advanced persistent threats early on its life cycle using prescriptive analytics. This research also propose some soft computing based general solutions to minimize the alert fatigue caused by generating too many false alerts by getting insights from security information and event management (SIEM) signals to make the prevention systems learn from the environment to keep false signals to a minimum