Stacking Ensemble Deep Learning for Real-Time Intrusion Detection in IoMT Environments

Area of Science:

• Cybersecurity in healthcare technology
• Network intrusion detection systems
• Machine learning applications in IoMT

Background:

• The Internet of Medical Things (IoMT) integrates interconnected devices for advanced patient care.
• IoMT systems handle sensitive data, making them vulnerable to cyber threats.
• Effective security solutions are crucial for IoMT network integrity.

Purpose of the Study:

• To develop a novel intrusion detection system (IDS) tailored for IoMT networks.
• To enhance cybersecurity by leveraging machine learning and deep learning techniques.
• To ensure real-time threat detection and classification in IoMT environments.

Main Methods:

• Implementation of a stacking ensemble method combining multiple machine learning and deep learning classifiers.
• Utilizing the Kappa Architecture framework for continuous processing of IoMT data streams.
• Developing an IDS capable of detecting and classifying various cyberattacks like ARP spoofing, DoS, Smurf, and Port Scan.

Main Results:

• The proposed IDS achieved high detection accuracy: 0.991 in binary classification and 0.993 in multi-class classification.
• The system demonstrated effective real-time performance through the Kappa Architecture.
• Successful detection and classification of diverse cyberattack types were achieved.

Conclusions:

• Combining advanced ML, DL, and ensemble learning offers a robust solution for IoMT cybersecurity.
• The developed IDS provides a reliable and scalable method for safeguarding healthcare services.
• This research addresses critical security challenges within the evolving IoMT landscape.