University of Twente Student Theses
Anomaly Detection for IoT : Another Look At Federated Learning
Bica, Ciprian (2024) Anomaly Detection for IoT : Another Look At Federated Learning.
|
PDF
1MB |
| Abstract: | As IoT devices integrate into daily life, they become indispensable for many different purposes. However, their increasing presence also makes them targets for malicious actors seeking to exploit any vulnerabilities. The motivations behind targeting IoT devices are diverse and often financially driven. For the most part, these target devices are not equipped with the best security measures to stay resilient against attackers, thus making them easy targets. One way of improving the security aspect of this system would be to analyse the network traffic of these devices and scan for and identify malicious attacks targeting them. Utilising the IoT-23 dataset, comprised of network traffic captures from various IoT devices, alongside a Federated Learning approach, the objective is to spot any anomalous traffic between these devices, which usually indicates an attack is happening. This study compares two established federated learning algorithms, FedAvg and FedProx, to determine their effectiveness in anomaly detection. Two different setups were tested, and it was found that for this dataset and the anomaly detection task, FedAvg seems to perform better in terms of accuracy, precision, recall and f-score on one of the setups, while on the other setup, the performance of the two algorithms was more similar. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 54 computer science |
| Programme: | Computer Science BSc (56964) |
| Link to this item: | https://purl.utwente.nl/essays/100788 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page