Application of Semantic Analysis and GMM Models for Anomaly Detection in Network Traffic

Abstract

Network traffic contains numerous patterns, and deviations from these patterns can indicate cyberattacks or system failures. Traditional machine learning methods, such as the Gaussian Mixture Model (GMM), are effective in detecting anomalies, but do not provide meaningful interpretations of these anomalies. This paper presents an approach that integrates semantic analysis with GMM to improve anomaly detection accuracy and provide contextual insights into abnormal behavior in network traffic. Using cybersecurity ontologies and semantic reasoning, detected anomalies can be mapped to known cyber threats, improving the reliability of detection. The proposed method is evaluated using real-world network traffic logs, demonstrating its effectiveness in reducing false positives and enhancing interpretability.