Analysis of methods for detecting network cyberattacks based on artificial intelligence

This thesis analyzes the methods of detecting and preventing network cyberattacks based on Artificial Intelligence (AI) and Machine Learning (ML) algorithms. It highlights that traditional security systems (firewalls, IDS) fail to provide sufficient defense against next-generation and complex hybrid attacks. The study explores and compares the capabilities of modern approaches such as Random Forest, SVM, and LSTM deep learning neural networks in anomaly detection. As a result, the high accuracy of AI models in detecting suspicious activities in network traffic in real-time is demonstrated, along with methods to mitigate the issue of false positives. Keywords: Cybersecurity, Artificial Intelligence, Machine Learning, Deep Learning, Network Traffic, Anomaly Detection, LSTM, Cyberattack Prevention, Real-time Analysis.