Research Article
Anomaly Detection in Cellular IoT with Machine Learning
@INPROCEEDINGS{10.1007/978-3-030-91421-9_5, author={Bernardo Santos and Imran Qayyrm Khan and Bruno Dzogovic and Boning Feng and Van Thuan Do and Niels Jacot and Thanh Van Do}, title={Anomaly Detection in Cellular IoT with Machine Learning}, proceedings={Smart Objects and Technologies for Social Good. 7th EAI International Conference, GOODTECHS 2021, Virtual Event, September 15--17, 2021, Proceedings}, proceedings_a={GOODTECHS}, year={2022}, month={1}, keywords={Machine learning Anomaly detection Mobile network security IoT security Cross layer security}, doi={10.1007/978-3-030-91421-9_5} }- Bernardo Santos
Imran Qayyrm Khan
Bruno Dzogovic
Boning Feng
Van Thuan Do
Niels Jacot
Thanh Van Do
Year: 2022
Anomaly Detection in Cellular IoT with Machine Learning
GOODTECHS
Springer
DOI: 10.1007/978-3-030-91421-9_5
Abstract
The number of Internet of Things (IoT) devices used in eldercare are increasing day by day and bringing big security challenges especially for health care organizations, IoT service providers and most seriously for the elderly users. Attackers launch many attacks using compromised IoT devices such as Distributed Denial of Services (DDoS), among others. To detect and prevent these types of attacks on IoT devices connected to the cellular network, it is essential to have a proper overview of the existing threats and vulnerabilities. The main objective of this work is to present and compare different machine learning algorithms for anomaly detection in the cellular IoT scenario. Five supervised machine learning algorithms, namely KNN, Naïve Bayes, Decision Tree and Logistic Regression are used and evaluated by their performance. We see that, for both normal (using a local test dataset) and attack traffic (CICDDoS2019 (CICDDoS2019 Dataset:https://www.unb.ca/cic/datasets/ddos-2019.html.)) datasets, the accuracy and precision of the models are in average above 90%.
