Résumé:
This dissertation focuses on the problem of attacks and intrusions into information systems, with particular emphasis on computer security and intrusion detection systems (IDS). It discusses and analyzes various intelligent classifiers used for automatic detection and classification of network attacks in IDSs. A significant part of the work focuses on ensemble feature selection, highlighting its advantages, potential challenges and best practices for effective use. In addition, it explores the basic models of IDS systems, their classification, available detection methods, as well as performance evaluation metrics. Finally, the brief guides readers in creating their own intrusion detection systems using the Weka tool and the NSL KDD dataset, to enhance network protection against cyber attacks.
