High Perception Intrusion Detection Systems Using Neural Networks

Computer networks threats are becoming one the most widely debated issues worldwide. Various types of attacks are generated periodically and in an increasing scale alarming and threatening networking security issues. This is leading to new serious and fast development of feasible techniques for developing effective Intrusion Detection Systems (IDS's). In this context, some of the widely used and tested IDS's are discussed. Anomaly-based network intrusion detection techniques are getting the attention in most of the networking fields to protect network systems against malicious acts, providing a good security level especially with the integration of neural networks in the detection systems which provided advanced methods of threats investigation and detection. This paper emphasizes the importance of anomaly-based intrusion detection techniques, the important outcomes of these systems, latest developed methods and what is expected from the future experiments in this field. Moreover, the technique of learning user profiles effects in detecting intrusions will be discussed. Finally, the lights will be shed on an offline approach using Multi Layer Perceptron (MLP) and Self Organizing Maps (SOM) which is a distinguished method in intrusion detection.