Publication

Abstract : The pervasive deployment of IoT devices has posed enormous security threats, highlighting the necessity for effective Intrusion Detection Systems (IDS) to safeguard networks. This study investigates the detection and classification of different types of attacks in IoT networks by utilizing a publicly available intrusion dataset and machine learning algorithms. The study compares five machine learning models: Elastic Net Regression, Random Forest, XGBoost, Neural Network Classifier, and Support Vector Machine (SVM). The models were applied to multiclass classification problems following extensive dataset pre-processing, including handling missing values, feature selection, and normalization. The models were tested using standard performance measures such as accuracy, precision, recall, and F1-score. XGBoost performed better than other methods, with 99.27% accuracy accompanied by 99% precision, recall, and F1-score (weighted average). Such remarkable findings confirm the immense potential of machine learning methods in enhancing IoT network security through advanced intrusion detection systems.