Real-Time Anomaly Detection in Network Traffic Using Graph Neural Networks and Random Forest

Network infrastructure security is a top issue in today's digitally linked world. The crucial issue of real-time anomaly identification in network data is addressed in this research study using Graph Neural Networks (GNNs) and Random Forest methods. Drawing on a sizable dataset obtained via honeypots put in various geographical regions, this research delves into a comprehensive investigation of how well these algorithms can foresee unexpected trends. The proposed strategy adheres to a rigid process that involves data preparation, model installation, thorough evaluation, and performance comparison. The study's conclusions about the relative benefits of Random Forest and GNNs in anomaly identification give significant new information. By leveraging visual tools including confusion matrices and anomaly score distributions, this study gives a complete view of the model outcomes. In this study, an actual data and a framework has been used for the optimal anomaly detection approach, boosting real-time network security.