Institutional Repository of Vilnius University
Detecting cyber attacks in real network traffic using dtw and machine learning /
Abstract
dc:descriptionIn this thesis, passive and near real-time methods are evaluated for cyber attack detection in real network traffic. A controlled infrastructure was built, integrating Apache Kafka with Apache Spark, and a custom dataset was created by generating realistic web-application attacks to be visible in network traffic and capturing it at the packet level. Passive analysis was performed with Dynamic Time Warping (DTW) to align and compare traffic sequences, which was informative when sequences were of similar length but less reliable when captures contained missing packets. For near real-time detection, packet data was aggregated into 5 second windows and scored using trained machine learning models. All evaluated models performed well, with Random Forest providing the best overall balance on the test set and practical deployment inside the Spark ML pipeline. The work shows that Kafka and Spark streaming can support timely detection, while DTW can be used to analyse network traffic passively for a deeper understanding.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Sipavičiūtė, Adrija,
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:ELABAETD277690261&prefLang=en_US