The article presents IoTGeM, a methodology for modelling IoT network attacks, focusing on generalizability and improved detection performance. It introduces an improved rolling window approach for feature extraction and a multi-step feature selection process that reduces overfitting. The models are built and tested using isolated datasets to avoid data leaks. The methodology is rigorously evaluated using diverse machine learning models, evaluation metrics, and datasets. The study also utilizes explainable AI techniques to identify the features that contribute to accurate attack detection.
Publication date: 4 Jan 2024
Project Page: Not provided
Paper: https://arxiv.org/pdf/2401.01343
