2026-05-05T14:37:10.654Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:3:11879 2026-04-13T23:21:29Z 3

Intrusion Detection Systems in IoT Networks by LightGBM Optimized with Altered Firefly Algorithm, Chapter in LNNS Lecture Notes in Networks and Systems: ICCCT 2025: International Conference on Communication and Computational Technologies, Springer, volume 1674 2026 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/3/11879 https://link.springer.com/chapter/10.1007/978-981-95-3498-2_26 N. Jankovic T. Zivkovic V. Markovic B. Radomirovic M. Zivkovic L. Jovanovic N. Bacanin Identification of intrusions plays a crucial role in Internet of Things (IoT) systems by protecting connected devices, networks, and data from access without authorization and ill-natured affairs. With IoT quickly expanding into areas like smart homes, autonomous cars, industry, and medical systems, ensuring the security of these devices became increasingly important since they are constrained in available resources and have diverse architectural models. This study tackles the intrusion detection challenge in IoT systems by employing the light gradient-boosting machine (LightGBM) classifier. A novel version of the widely renowned firefly algorithm was introduced for fine-tuning of the LightGBM classifier’s hyperparameters, leveraging its capabilities in detecting intrusions in IoT systems. Comprehensive comparative simulations were executed, benchmarking the suggested approach against a range of other robust optimizers in identical setup. The simulation outcomes highlighted the distinguished performance attained by the suggested model, demonstrating its considerable prospect within this particular area. bookPart Springer, Singapore 341 360 10.1007/978-981-95-3498-2_26 LNNS Lecture Notes in Networks and Systems: ICCCT 2025: Proceedings of International Conference on Communication and Computational Technologies, volume 1674