Answers are grounded in this paper and cite their sources.
Network intrusion detection requires the identification of malicious traffic while maintaining the ability to observe unusual network behavior. This paper presents Dynamic Cybersecurity Sentinel, a network monitoring and intrusion detection system that combines supervised machine learning with behavioral anomaly detection. The system uses the UNSW-NB15 dataset for the development and evaluation of its classification component. An XGBoost-based classifier is used to categorize network traffic into Normal, Reconnaissance, Exploits, and Generic classes. In addition to classification, the system incorporates Isolation Forest for behavioral anomaly detection. The live monitoring part uses the Scapy library to catch network packets, which are then turned into network flows before extracting and analyzing features. The information from this process goes through baseline calibration, risk and severity evaluation, recommendation for responses, storing alerts, and showing results on a dashboard. The system was built in two steps, starting with a structured CSV-based setup and moving on to real-time network traffic monitoring. Testing the final XGBoost classifier showed an accuracy of 75.00%, weighted precision of 0.83, weighted recall of 0.75, and weighted F1-score of 0.77.The live version was tested using traffic from several connected computers. The results show how classification, behavior analysis, risk assessment, alert creation, and visualization work together in one network security monitoring system
No takes yet. Share an insight, caveat, or question.
Wanjari et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: