A Concept for Minimizing False Alarms and Security Compromise by Coupled Dynamic Learning of System with Fuzzy Logics

Objectives: To develop a novel method of Intrusion Detection System (IDS) by coupled dynamic learning of system with Fuzzy logics for minimizing false alarms and security compromise of a system connected with internet. Method: When Intrusion Detection System (IDS) raise alarm based on assigned rules, there would be a possibility for too many false alarms. The degree of intrusion and subsequent alert are often depending on different situations. These situations are not unique for all systems hence; a global knowledge based filter rules fail to minimize false alarms. In this paper, a concept was proposed to solve this hazy and unclear cutoff rules derived from global knowledge, by self-learning and turning activity of system, towards the security issues from the analytical outcomes of behavioral patterns of network system. Findings: The use of fuzzy logic helps to smooth the sharp separation of normal and abnormal behaviors in network activity which adds further strength in minimizing false alarms and security compromise. This concept is illustrated and demonstrated using some familiar network behaviors for easy understanding of logics and mechanism of the proposed IDS model. Application/Improvements: This intelligence associated with fuzzy logic may be extended with more and more parameters for better efficiency in Intrusion Detection System (IDS).