A hybrid XGBoost-SVM ensemble framework for robust cyber-attack detection in the internet of medical things (IoMT).

Today, the rise of the Internet of Medical Things (IoMT) has evolved into a highly valued global market worth billions of dollars. However, this growth has also created many opportunities for massive and advanced attack scenarios due to the vast number of devices and their interconnected communication networks. Based on recent reports, it is observed that during the Covid-19 pandemic, the necessity of the IoMT ecosystem has increased significantly. On the other hand, attackers and intruders aim to impair data integrity and patient safety with the prevalence of sophisticated cyber attacks including Man in the Middle (MITM) attacks like spoofing and data injection. In this research work, WUSTL-EHMS-2020 dataset is utilized to demonstrate a robust IoMT cyberattack detection method based on machine learning and the efficiency of the proposed model is validated by employing TON-IoT and CICIDS 2017 datasets. We offer an ensemble approach that employs Extreme Gradient Boosting (XGBoost) and Support Vector Machine (SVM) classifiers to address the exclusive challenges in IoMT security. By utilizing the complementing advantages of SVM's decision boundary precision and XGBoost's gradient-based optimization, our model outperforms baseline techniques with a superior detection accuracy of 98.04% with WUSTL-EHMS-2020 dataset. In addition, the practicality of the proposed model is examined by considering peculiar features of IoMT like resource restrictions, medical device communication diversity, and healthcare data privacy by comparing with IoMT and IoT datasets by exposing the patterns of cyberattack in dynamic IoMT environment. Hence, this research will be considered as the pioneer for developing reliable IoMT security solutions by adapting trustworthy and scalability.