Relevant Feature Selection Model Using Data Mining for Intrusion Detection System

Network intrusions have become a significant threat in recent years as a result of the increased demand of computer networks for critical systems. Intrusion detection system (IDS) has been widely deployed as a defense measure for computer networks. Features extracted from network traffic can be used as sign to detect anomalies. However with the huge amount of network traffic, collected data contains irrelevant and redundant features that affect the detection rate of the IDS, consumes high amount of system resources, and slowdown the training and testing process of the IDS. In this paper, a new feature selection model is proposed; this model can effectively select the most relevant features for intrusion detection. Our goal is to build a lightweight intrusion detection system by using a reduced features set. Deleting irrelevant and redundant features helps to build a faster training and testing process, to have less resource consumption as well as to maintain high detection rates. The effectiveness and the feasibility of our feature selection model were verified by several experiments on KDD intrusion detection dataset. The experimental results strongly showed that our model is not only able to yield high detection rates but also to speed up the detection process. Keywords— Intrusion detection system, traffic classification, network security, supervised learning, feature selection, data mining.