Abstract
This study examines the effectiveness of applying artificial intelligence (AI) technologies to detect anomalies in network traffic for improving information security. The rapid growth of digital infrastructures, cloud computing, and Internet-connected devices has significantly increased the volume and complexity of network traffic, making conventional signature- and rule-based detection methods less effective against modern cyber threats. As a result, intelligent approaches capable of identifying previously unknown attack patterns have become essential for ensuring the security and reliability of computer networks. The research investigates the application of machine learning and deep learning techniques for network traffic anomaly detection. Particular attention is given to the comparative analysis of widely used algorithms, including Random Forest and Long Short-Term Memory (LSTM) neural networks, with respect to their detection accuracy, computational efficiency, and ability to recognize complex temporal patterns in network traffic. The proposed approach also considers the integration of synthetic and real-world datasets to improve the reliability of model training and validation under diverse operating conditions. The findings demonstrate that AI-based models provide significantly higher detection accuracy, lower false-positive rates, and greater adaptability than traditional approaches. Among the evaluated methods, LSTM showed superior performance in identifying temporal anomalies, whereas Random Forest achieved a favorable balance between accuracy and computational cost. The obtained results confirm the potential of artificial intelligence for real-time network monitoring, early detection of cyberattacks, and automated incident response. The study concludes with practical recommendations for integrating AI-based anomaly detection systems into modern cybersecurity infrastructures to enhance the resilience, reliability, and overall effectiveness of network protection mechanisms.