AI AGAINST INTRUSIONS: MACHINE LEARNING APPROACHES FOR CYBER ATTACK DETECTION

Abstract

The rapid growth of digital technologies and interconnected systems has led to an alarming rise in cyber threats, posing significant risks to network security, data privacy, and organizational resilience. Traditional rule-based intrusion detection systems often fail to identify novel or sophisticated attacks, creating the need for intelligent and adaptive solutions. Machine learning (ML) techniques offer a powerful alternative by learning patterns of normal and malicious network behavior, enabling early detection and prevention of cyber-attacks. This study reviews the application of ML algorithms—including decision trees, support vector machines, random forests, k-nearest neighbors, and deep learning models—in detecting various forms of network intrusions such as denial-of-service (DoS), malware injection, phishing, and advanced persistent threats. Emphasis is placed on feature selection, classification accuracy, anomaly detection, and the challenges of imbalanced datasets. Experimental evidence demonstrates that MLdriven approaches significantly improve detection rates, reduce false alarms, and adapt better to evolving attack vectors compared to conventional systems. The review highlights machine learning as a cornerstone in the development of next-generation, intelligent cybersecurity frameworks.