Network Intrusion Detection by Artificial Immune System and Neural Network

Abstract

Easy access, simulation of IOT network increases its application and demands in different area. As many of IOT networks are vulnerable in nature and attracts intruders to take advantage of weak security. This paper has developed a model that can detect the IOT network intrusion. In this work feature optimization was done by use of artificial immune  system algorithm. AIS reduces the dimension of the dataset by applying affinity check and cloning steps. Selected features were further use for the traiing of neural network. Trained neural network predict the class of IOT network session (Normal / Malicious). Experiment was done on real dataset of IOT session and result shows that rpopsoed model has improved the detection accuracy as compared o existing models.