Improved security intrusion detection using intelligent techniques

Nowadays, the information systems security is a crucial issue for the survival of any company, so this justifies the use of intrusion detection systems (IDS) or the intrusion prevention systems (IPS). These systems are essentially based on the analysis of the network data content (frames), in search of traces of known attacks. Currently, IDS/IPS become the main element of security networks and hosts, they can both detect and respond to an attack in real time or off-line. Even this, having a completely secure network is practically impossible. In this article, we try to propose an improvement of intrusion detection systems based on Machine Learning techniques. These rapidly expanding techniques have shown that predictions and machine learning could be improved, which could significantly improve the reliability of detection against polymorphic and unknown threats. Simulation results showed that security intrusion detection is improved with the use of Machine Learning techniques.