Conceptual framework and theoretical challenges in unsupervised anomaly detection for network traffic data

Abstract

This paper outlines the initial design considerations for a system aimed at processing and analyzing network traffic data to detect anomalies using machine learning. The study explores anticipated challenges in data preprocessing, scalability, and algorithm selection, emphasizing the potential of unsupervised learning methods to identify unusual patterns in network traffic. The proposed approach serves as a foundation for future development of anomaly detection systems.