Performance Analysis of Artificial Immune System With Skewed Class Data

Abstract

Artificial Immune Systems (AIS) are algorithms inspired by the human immune system. The human immune system is a highly adaptable and reliable system and is extremely complex in nature. Neural Network and Genetic principles based solutions have been successfully deployed; however it performs very well in systems with low standard deviation between the class labels.  In this paper we investigate AIS algorithms for network intrusion classification which is characterized by skewed class set with true positive being extremely low.