NETWORK SECURITY ENHANCEMENT THROUGH MACHINE LEARNING-BASED CYBER ATTACK DETECTION

Abstract

The objective of this investigation is to identify and prevent harmful activities in contemporary communication networks by employing machine learning-based techniques to enhance network security and detect intrusions. Using supervised and unsupervised machine learning algorithms, including Decision Trees, Random Forest, Support Vector Machines, and Neural Networks, the investigation examines network traffic patterns and identifies outliers that may be the result of cyber threats such as phishing, DOS attacks, malware intrusions, and unauthorized access attempts. The proposed method analyzes a variety of network metrics, including packet flow, connection duration, source and destination IP behavior, and protocol utilization, to facilitate the identification of both normal and anomalous activity. Traditional security systems that rely on signatures are significantly outperformed by machine learning-driven detection models in terms of real-time attack detection, false alarm rates, and threat detection speed. The paper underscores the critical nature of enhanced cyber security systems in order to ensure the security of private information, the reliability of networks, and the safety of online communication.