Institutional Repository of Vilnius University
Predictive modelling for cyber threat intelligence /
Abstract
dc:descriptionThis thesis addresses the problem of real-time detection and prevention of network attacks. Traditional security systems that rely solely on known attack signatures often miss new threats. The paper explores how machine learning can help identify suspicious patterns in network traffic, testing various algorithms such as XGBoost, Random Forest, and logistic regression to classify network flows as normal or malicious. This research led to the development of "IronFlux IDS", a comprehensive intrusion detection system that combines machine learning-based classification with rule-based heuristics and log analysis. The system intercepts network packets in real time, analyzes them using trained machine learning models, and automatically blocks suspicious IP addresses. It includes a web dashboard for monitoring attacks, viewing statistics, and managing threats. The system is built on Python using FastAPI for the backend, React for the user interface, and MySQL for data storage, all running in Docker containers. Testing shows that the system successfully detects various types of attacks, including port scanning, DDoS attacks, password brute-force attempts, and web vulnerabilities. Machine learning is more effective than traditional methods in detecting new attack patterns, and combining multiple detection approaches increases accuracy while reducing false positives. The implemented solution demonstrates that automatic threat detection and prevention can be effective in real-world network environments.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Anapolskyi, Andrii,
Subjects
dc:subject × 1Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- eng
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD277689899&prefLang=en_US