Application of Artificial Immune System Algorithms in Anomaly Detection

Anomaly detection is a concept which is widely applied to a number of domains. Anomaly-based intrusion detection system (ABIDS) has the ability to detect previously unknown attacks which are important since new vulnerabilities and attacks are constantly appearing. There are several techniques of anomaly detection techniques which are developed over the years, practically and in research area to observe new attacks on the network systems. This paper proposes the better artificial immune system (AIS) algorithm for anomaly detection. We have taken two datasets; one is the NSL-KDD dataset which purely focuses on intrusion detection (ID) and another Adult dataset which is not meant for ID. Here, we have performed our assessment by using the NSL-KDD dataset which is an altered version of the broadly utilized KDDcup99 dataset. The evaluation for selection of better AIS algorithm is done by several parameters such as accuracy, specificity, False Negative Rate, False discovery Rate, and Negative Predicted Value.