AN INTRUSION DETECTION SYSTEM BASED ON AN ENSEMBLE-BASED FEATURE SELECTION PROCESS

Alshaibi Ahmed Jamal, Al-Ani Mustafa Majid, Al-Alazzawi Abeer Yassin, Konev Anton Alexandrovich

The article analyzes software that provides security monitoring and real-time threat detection. It is shown that most of the methods have a large intrusion detection time due to the high traffic noise. A new method for selecting intrusion signs based on obtaining information, its classification and creating options using the particle swarm method is proposed. This method consists in a preliminary selection of sample groups characteristic of various invasion methods. Based on this method, a technological scheme for intrusion detection and a platform for pre-configuring an intrusion detection detector for the main groups of intrusions and attacks are proposed. Experimental data comparing the proposed method with the reference method without assembly showed its efficiency almost two times higher. In conclusion, it is noted that the possibility of using this method in the field of cloud computing. The high classification accuracy of ensemble classifiers compared to a single machine learning algorithm for detecting IDS in the cloud shows great promise for this method.

Keywords:information protection, security, monitoring, intrusion detection, network anomalies, anomaly assembly