Feature efficiency in IoMT security: A comprehensive framework for threat detection with DNN and ML

Area of Science:

• Cybersecurity
• Healthcare Technology
• Machine Learning

Background:

• Critical security challenges in the Internet of Medical Things (IoMT) require advanced solutions.
• Existing IoMT cybersecurity measures often face limitations in detection accuracy and computational efficiency.
• Optimizing feature selection is crucial for enhancing the security and integrity of real-time healthcare systems.

Purpose of the Study:

• To develop a feature selection framework for IoMT cybersecurity.
• To improve detection accuracy and computational efficiency in IoMT security.
• To enhance the security and operational integrity of real-time healthcare systems.

Main Methods:

• Integration of Random Subset Feature Selection (RSFS) with Correlation Feature Selection (CFS).
• Development of a novel feature selection framework tailored for IoMT datasets.
• Testing on four diverse IoMT datasets using multiple machine learning models (Random Forest, KNN, SVM, XGBoost, DNN).

Main Results:

• Exceptional detection accuracies achieved: 99.82% (TON-IoT), 99.99% (ICU-Dataset), 96.37% (WUSTL-EHMS-2020), and 99.99% (ECU-IoHT).
• Demonstrated significant improvements over existing methods with a reduced feature set.
• Showcased enhanced detection accuracy and processing efficiency for high-dimensional IoMT data.

Conclusions:

• Introduced a robust and scalable feature selection framework for IoMT cybersecurity.
• Provides a practical solution to security gaps, enhancing patient data protection and operational resilience.
• Holds potential for broad implementation in safeguarding critical IoMT infrastructures and advancing secure healthcare systems.