An Intelligent Intrusion Detection Model for MANET’s based on Hybrid Feature Selection

Intrusion Detection Systems (IDS) use different data reduction techniques. These techniques influence to improve the performance of the system through detect the attackers. At the same time the existing techniques posses slow training and testing process and are more expensive as well. Moreover, the false alarm rate in such system is high. Therefore, preprocessing is an important issue in Intrusion Detection System. In this paper, we propose a new hybrid attribute selection technique based on Extended Chi-square and Enhanced Multiclass Support Vector Machine (EMSVM) algorithm to build an Intrusion Detection System. Verification tests have been carried out by using the KDD’99 Cup Dataset. From the experiments, it is observed that significant improvement has been achieved from the viewpoint of both high detection and low false alarm rate.