Performance Evaluation of Artificial Immune System Algorithms for Intrusion Detection using NSL-KDD and CICIDS 2017 Datasets

Abstract

Artificial Immune System (AIS) algorithms are used to build models for classification and some clustering problems if there is availability of an effective dataset. A dataset that contains benign and common attack network flows that mimics the real time can only help to train and test an intrusion detection system. In this paper, classification models for anomaly-based intrusion detection are built using AIS algorithms namely AIRS1, Immunos1 and ClonalG. These algorithms were tested with NSL-KDD and CICIDS 2017 datasets, which have common updated set of malicious attacks such as DDoS, XSS, SQL Injection and Botnet. Our experiments show that AIS algorithms performs better in detecting new attacks than other classifiers. The outcome of this research has improved intrusion detection system by testing for attack diversity.