Enhancing IoMT network security using ensemble learning-based intrusion detection systems

An Intrusion Detection System (IDS) is crucial for protecting Internet of Medical Things (IoMT) networks by detecting malicious traffic. This paper examines the use of ensemble learning to enhance IDS effectiveness in IoMT environments. We employ base models including Random Forest (RF), XGBoost (XGB), AdaBoost (ADA), Multi-Layer Perceptron (MLP), Logistic Regression (LR), and K-Nearest Neighbors (KNN). The performance of the integrated ensemble model is evaluated using metrics such as accuracy, precision, recall, and F1-score. Our results demonstrate that the weighted voting ensemble of LR, ADA, and XGB achieved high performance, completing tasks in 12.42 seconds with an accuracy of 0.9960. The confusion matrix results further confirm its efficacy in accurate classification.