An efficient intrusion detection mechanism based on particle swarm optimization and KNN

In this paper an efficient intrusion detection mechanism based on particle swarm optimization and KNN has been presented. In our approach experimentation has been performed for the intrusion detection considering NSL-KDD dataset. Then the selected weights have been added directly to the final classification which has been received safely. Then the remaining selected weights have been added for the classification. These nodes are originally safe but received unsafe. It has been input for the classification process. KNN has been used for the classification of the initial features and the content features. The remaining features have been transferred to the particle swarm optimization. PSO has been used for the classification of the traffic and host features. It has been classified based on 50% threshold value. The results show that by using our approach the average classification accuracy is approximately 98%. The attack considered here are Denial of Service (DoS), User to Root (U2R), Remote to User /Login (R2L) and Probe.