Layered Approach For Intrusion Detection Using Multiobjective Particle Swarm Optimization

Intrusion detection is one of the challenging tasks in today’s networked world. It is necessary to formulate a new intrusion detection system, which can monitor the network to detect the malicious activities. The proposed work focuses the issues, namely accuracy and efficiency. One way to improve performance is to use a minimal number of features to define a model in a way that it can be used to accurately discriminate normal from anomalous behavior. So the new system uses an optimized feature selection algorithm to produce the reduced set of features and high attack detection accuracy can be achieved by using a layered approach. The feature selection algorithm used in the proposed system is a multi-objective particle swarm optimization algorithm which does the feature selection effectively. The layered approach is effectively applicable to detect anomaly attack. The proposed system is tested with the benchmark KDD ’99 intrusion data set as well real time captured data set, which outperforms other well-known methods such as the decision trees, naive Bayes and Ant Colony optimization. The system is highly robust and efficient. It can deal with real-time attacks and detect them fast and quick response.