Answers are grounded in this paper and cite their sources.
Evaluates hybrid DL IDS framework in IoT environments, suggesting effective cyberattack defense.
The rapid growth of Internet of Things (IoT) ecosystems has increased exposure to diverse cyberattacks, necessitating accurate and resource efficient Intrusion Detection Systems (IDS). This article proposed a hybrid Deep Learning (DL) IDS framework that integrates Bidirectional Gated Recurrent Units (BiGRU) with two metaheuristic optimizers: Improved Wild Horse Optimizer Algorithm (IWHOA) for selecting the most informative features and the Fennec Fox Algorithm (FFA) for tuning BiGRU parameters. The model is evaluated on three benchmark datasets ToN-IoT, NSL-KDD, and CIC-IDS2018 suing stratified splitting and 5-fold cross-validation. The proposed IDS achieves up to 99.18% accuracy, 98–99% average F1-score, and reduces feature dimensionality by 32–47% through IWHOA. Minority attack classes show major improvements, with F1-score gains of 4–7% compared to state-of-the-art baselines. The false positive rate remains consistently below 1.2% across all datasets. Inference tests on IoT hardware demonstrate practical deployability, with model size at 1.7 MB and average inference latencies of 42 ms on Raspberry Pi 4 and 18 ms on Jetson Nano. Overall, the findings confirm that the BiGRU-IWHOA-FFA hybrid system provides a high accuracy, low footprint and generalizable IDS solution suitable for real time IoT security.
No takes yet. Share an insight, caveat, or question.
Sriramappa et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: