Early Detection of Botnets Using Artificial Intelligence Methods

Computer systems worldwide face various threats daily, with botnets being one of the most serious. Detecting botnet activity is challenging due to the need to monitor vast amounts of heterogeneous traffic. As bots’ architecture and communication methods evolve, the detection process becomes increasingly complex. This paper proposes an approach to detecting botnet activity by identifying network traffic anomalies using machine learning. It also examines the main existing types of botnets, classifying them based on the purpose of the attack and the type of control. Furthermore, it identifies the primary modern methods for detecting botnet activity. The research concludes that existing detection methods may not be sufficiently effective as botnets continue to evolve. Detection experiments were conducted using various machine learning algorithms on recent datasets to determine which algorithms are most effective in detecting botnet activity. This paper will be valuable for developers of intrusion detection systems and researchers in the field of computer security.