Neural Network Based Intrusion Detection: State of the Art

Abstract

With the advancement in technology and over internet, the number of attacks through unauthorized access has also increased. For fighting these attacks and to ensure the safety of the system, powerful Intrusion Detection System (IDS) is required. IDS’s are used for sensing the attack persisting inside the network system. This paper reviews different existing Intrusion Detection Systems using Artificial Neural Networks for detecting malicious network activity as the success of neural networks comes from the fact that they are able to learn a number of behaviours depending on network input. They are very effective and fast in classification and can easily identify new threats. One of the features that uphold neural networks in the field of intrusion detection is its flexibility to adapt to any environment. This work evaluates the parameters that play the major role in enhancing the efficiency and accuracy of IDS.