Electronic Thesis/Dissertation
 

A Machine Learning Intrusion Detection System (IDS) Tool for Healthcare Internet of Things (IoT) Devices

Open Access Deposited

Abstract of PraxisA Machine Learning Intrusion Detection System (IDS) Tool for Healthcare Internet of Things (IoT) Devices The rapid expansion of Internet of Things (IoT) and Industrial Internet of Things (IIoT) devices in healthcare has introduced numerous security challenges. Cyberattacks including Distributed Denial of Service (DDoS), IoT reconnaissance, Man-in-the-Middle (MITM) and injection attacks, and other malware threats have surged among devices with diverse protocols and limited computing power. Intrusion Detection Systems (IDSs) play a pivotal role in safeguarding communication across IoT devices and fortifying their security posture. In recent years, various machine learning (ML) techniques have been widely adopted for IDSs in IIoT systems, yet there remain areas for improvement, particularly in building more trustworthy and interpretable systems, as well as handling anomaly detection within dynamic environments characterized by shifting behaviors over time. The proposed novel ML-based IDS tool for IIoT devices is a robust, trustworthy, and interpretable IDS that integrates SHapley Additive exPlanations (SHAP) values and leverages eXtreme Gradient Boosting (XGBoost) and Random Forest (RF) learning techniques. This ML-IDS tool incorporates K-10 cross-validation to ensure generalization ability and reliability to dynamically identify cyberattacks on IoT healthcare systems. Its performance metrics exceed 99.95% accuracy, 99.96% precision, 99.96% recall, and 99.96% F1-score in multi-class classification, and achieved 100% accuracy in binary classifications with minimal false positives or false negatives. Furthermore, consistent Receiver Operating Characteristic (ROC) and Area Under the Curve (AUC) metrics demonstrate 100% accuracy in both binary and multi-class classifications, affirming the robustness and validity of the ML-IDS approach.

Author Language Keyword Date created Type of Work License
  • All rights reserved
Rights statement GW Unit Degree Advisor Committee Member(s) Persistent URL

Notice to Authors

If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.

Thumbnail Title Date Uploaded Visibility Actions
Preview of Alemu_gwu_0075A_17118.pdf Alemu_gwu_0075A_17118.pdf 2025-04-09 Open Access