Anomaly Based Intrusion Detection using Feature Relevance and Negative Selection Algorithm

With the increase in the use of internet, the job of malicious people has been made easy to exploit vulnerabilities in existing system. Intrusion Detection System (IDS) plays a major role in computer/network security in recognizing such malicious activity called intrusion. IDSs’ quick and correct detection of unknown intrusions help in reducing damage/loss to the sensitive information. IDS with such quality need to be encouraged. An IDS must analyse a different set of attributes to decide upon a monitored information as normal or abnormal. This becomes a time consuming task if the number of attributes to be analysed are more. Hence this becomes a barrier for fast detection. This paper proposes and implements an intrusion detection approach that breaks this barrier and helps in correct detection of novel attacks. This is achieved by using significant feature set extracted from KDDcup1999 dataset and Negative Selection Algorithm (NSA) an Artificial Immune System concept.