Supplementary material - An Optimized Gradient Boosting Framework for IoT Intrusion Detection: A Comprehensive Evaluation on the CIC-IoT-2023 Dataset

This repository provides the complete supplementary material for the study “An Optimized Gradient Boosting Framework for IoT Intrusion Detection: A Comprehensive Evaluation on the CIC-IoT-2023 Dataset.”

It includes three preprocessed feature-selected datasets — Binary, Eight-Class, and Thirty-Four-Class — derived from the CICIoT2023 benchmark using stratified undersampling and LightGBM-based feature selection. These datasets contain only the discriminative features used in the experiments and are ready for direct model reproduction without further preprocessing.

Each classification level is accompanied by:

A comprehensive Excel results file (*_Classification_FullResults.xlsx) summarizing class distributions, feature importance, model performance reports, confusion matrices, inference latency, and overall accuracy/F1 summaries.

A reproducible experiment script (*_Classification_Experiment.py) that automates the entire workflow — including data loading, normalization, model training (XGBoost, LightGBM, CatBoost), and evaluation.

All resources are released under the Creative Commons Attribution 4.0 International License (CC BY 4.0).Original data were obtained from the CICIoT2023 dataset by the Canadian Institute for Cybersecurity (UNB).For full documentation and experiment notebooks, please visit the companion GitHub repository:🔗 GitHub Repo