Anomaly Detection for the Internet-of-Medical-Things

Cybersecurity for the Internet of Medical Things (IoMT) is a very concerning issue because of emerging cyber threats and security incidents targeting IoMT devices all over the world. The healthcare system has near-zero tolerance for inexplicability. In this paper, we propose a machine learning-based anomaly detection for the IoMT and evaluate the performance using a realistic public dataset. We implement various machine learning algorithms: Random Forest, Decision Tree, Logistic Regression, Support Vector Machine, and K-Nearest Neighbor with TON_IoT dataset. Two types of classifications are implemented: binary and categorical. In the categorical classification, evaluation for nine attack scenarios (Scanning, DoS, password cracking attack, and Man-in-The-Middle (MITM)) are performed. The test results demonstrate that Support Vector Machine models produce better performance compared to the other models.