Scholarlink

Purpose <p>The Internet of Health Things (IoHT) is a precise adaptation of the Internet of Things (IoT) in the health domain that allows medical objects to be embedded with electronic and networking capabilities for real-time medical data exchange. Their increased adoption comes at the expense of widening cyberattack surfaces and other unintended cybersecurity consequences that require sophistication to address. This paper aims to employ AI and ML techniques to strengthen cybersecurity practices in the healthcare sector.</p> Methods <p>For our methods, we used a streamlined and adapted ML pipeline on the Edith Cowan University (ECU) IoHT; a dataset developed primarily for the analysis and evaluation of network traffic. Furthermore, we compared several approaches employed in anomaly detection in IoHT environment and converged with four classification AI/ML techniques of Gradient Boosting (GB), Decision Trees (DT), Random Forest (RF) and Multi-Layer Perceptron (MLP).</p> Results <p>A comprehensive comparative analysis is conducted based on key performance metrics such as accuracy, precision, recall, and F1-score. The results showed impressive classification accuracy of more than 90% in classifying ARP spoofing, DoS, Nmap port scan and smurf attack types. Experimental results demonstrate that each algorithm exhibits unique strengths in different aspects of IoHT security.</p> Conclusions <p>In conclusion, the findings of this study provide valuable insights into selecting appropriate AI algorithms for specific IoHT security requirements, contributing to the development of more resilient and effective security mechanisms in healthcare systems leveraging IoHT technologies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Maximizing healthcare security outcomes through AI/ML multi-label classification approach on IoHT devices

  • 期刊论文

  • 发表日期：

    2025年4月8日

查看全文
  • Emmanuel Song Shombot,
  • Gilles Dusserre,
  • Robert Bestak,
  • Nasir Baba Ahmed

摘要

Purpose

The Internet of Health Things (IoHT) is a precise adaptation of the Internet of Things (IoT) in the health domain that allows medical objects to be embedded with electronic and networking capabilities for real-time medical data exchange. Their increased adoption comes at the expense of widening cyberattack surfaces and other unintended cybersecurity consequences that require sophistication to address. This paper aims to employ AI and ML techniques to strengthen cybersecurity practices in the healthcare sector.

Methods

For our methods, we used a streamlined and adapted ML pipeline on the Edith Cowan University (ECU) IoHT; a dataset developed primarily for the analysis and evaluation of network traffic. Furthermore, we compared several approaches employed in anomaly detection in IoHT environment and converged with four classification AI/ML techniques of Gradient Boosting (GB), Decision Trees (DT), Random Forest (RF) and Multi-Layer Perceptron (MLP).

Results

A comprehensive comparative analysis is conducted based on key performance metrics such as accuracy, precision, recall, and F1-score. The results showed impressive classification accuracy of more than 90% in classifying ARP spoofing, DoS, Nmap port scan and smurf attack types. Experimental results demonstrate that each algorithm exhibits unique strengths in different aspects of IoHT security.

Conclusions

In conclusion, the findings of this study provide valuable insights into selecting appropriate AI algorithms for specific IoHT security requirements, contributing to the development of more resilient and effective security mechanisms in healthcare systems leveraging IoHT technologies.