Validating a real 5G SA jamming dataset through supervised learning baselines
Journal of Logic and Computation · 02.11.2026 · 36 (8)
DOI: 10.1093/logcom/exag056 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Logic and Computation, online erschienen am 13.11.2026. Verfasst von Diego Narciandi-Rodríguez und 5 weiteren Autor:innen. Die Arbeit nennt 49 Literaturangaben.
Abstract (Originalsprache)
Private 5G Standalone (SA) networks face radio-frequency jamming as an availability threat that propagates from the radio interface into observable changes in control-plane and user-plane traffic. This work assesses whether a labelled 5G SA dataset, captured in a private Firecell testbed under controlled software-defined-radio-based jamming and including legitimate disconnection events labelled as normal, supports supervised jamming detection from protocol-derived features. Four classical baselines (Logistic Regression, Gradient Boosting, k-Nearest Neighbours and Decision Tree) are trained on a stratified split with grid-search hyperparameter tuning, after tabular preprocessing of the traces. The evaluation is repeated over five random partitions and reported as mean and standard deviation. The dataset is also compared with existing 5G intrusion detection datasets and characterized at the protocol-stack level. Gradient Boosting yields the best trade-off and the highest PR-AUC, with the smallest dispersion across partitions, while Logistic Regression gives the highest jammer recall at a high false-positive rate and k-Nearest Neighbours proves the most sensitive to the partition. The dataset is a realistic basis for subsequent work on more advanced detection models.
Zitieren
APA
Narciandi-Rodríguez, D., Martínez-Martínez, G., Aveleira-Mata, J., Bayón-Gutiérrez, M., Alfonso-Cendón, J., & García-Rodríguez, I. (2026). Validating a real 5G SA jamming dataset through supervised learning baselines. Journal of Logic and Computation, 36(8). https://doi.org/10.1093/logcom/exag056
Vancouver
Narciandi-Rodríguez D, Martínez-Martínez G, Aveleira-Mata J, Bayón-Gutiérrez M, Alfonso-Cendón J, García-Rodríguez I. Validating a real 5G SA jamming dataset through supervised learning baselines. Journal of Logic and Computation. 2026;36(8). doi:10.1093/logcom/exag056
BibTeX
@article{NarciandiRodrguez2026,
author = {Narciandi-Rodríguez, Diego and Martínez-Martínez, Guillermo and Aveleira-Mata, Jose and Bayón-Gutiérrez, Martín and Alfonso-Cendón, Javier and García-Rodríguez, Isaías},
title = {Validating a real 5G SA jamming dataset through supervised learning baselines},
journal = {Journal of Logic and Computation},
year = {2026},
volume = {36},
number = {8},
doi = {10.1093/logcom/exag056}
}
EUVIMED