In this paper, we present an intrusion detection system which exploits pattern recognition techniques to model the usage patterns of authenticated users and uses it to detect intrusions in wireless networks. The key idea behind the proposed intrusion detection system is the identification of discriminative features from users activity data and use them to identify intrusions in wireless networks. The detection module uses PCA technique to accumulate interested statistical variables and compares them with the thresholds derived from users activities data. When the variables exceed the estimated thresholds, an alarm is raised to alert about a possible intrusion in the network. The novelty of the proposed system lies in its light-weight design which requires less processing and memory resources and it can be used in real-time environment.
展开▼
