Work

Detection of network attacks in internet of things environments through traffic visualization and analysis with deep learning

Year: 2026
Type: dissertation
Abstract: This thesis focuses on the design and implementation of an advanced anomaly detection system in IoT (Internet Of Things) networks, combining machine learning techniques, securing an IoT network using Explainable Artificial Intelligence (Explainable AI -XAI), utilizing multi-level data enrichment, analyzing the necessary packets, creating PCAP log files.At the same time, a PCA analysis was performed, an Isolation Forest algorithm was implemented, while a simultaneous attack on the network was performed in order to analyze suspicious packets as well as all network traffic.In addition, an analysis was performed using the Z scores type, various types of network attacks were analyzed, data extraction and collection were performed, Suricata was used and the necessary graphs -diagrams were extracted which demonstrate all the results. 2 VPNs were created for the purpose of remote management -supervision of the network.In summary, unsupervised machine learning algorithms in collaboration with IDS, IPS systems contributed greatly.In the end, the manual network security process and the basic concept of the rules of a properly configured firewall are shown, while its dynamics in real time are shown.
Language: Greek
Open Access status: green
OpenAlex ID: W7212348733
Cites: 32
Cited by:
Related to: 10