Implementing Stratified K-Fold Technique for Data Intrusion Detection Models in Internet of Health Things Systems

Simple item page
dc.contributor.authorOsa Edosa
dc.contributor.authorIyekekpolo Enosa N.
dc.contributor.authorIbhaze Augustus E.
dc.contributor.authorSadoh Wilson E.
dc.contributor.authorOrukpe Patience E.
dc.date.accessioned2024-11-26T09:53:59Z
dc.date.available2024-11-26T09:53:59Z
dc.date.issued2024-11-25
dc.description.abstractDue to the contemporary prevalence of evolving cutting-edge computerized technologies in virtually every sphere of modern day human activities, many of such interventions find their use in the medical domain. One such technology that is widely adopted is the Internet of Things, which when implemented in the medical industry is dubbed Internet of Health Things (IoHT) or Internet of Medical Things (IoMT). Such systems are widely adopted due to the increased mobility and dynamism they introduce to the dispensing of health services. They therefore are repositories of a lot of medical data, be it patient records, modus operandi of sensitive and competitive healthcare service delivery, geolocation records for disease outbreaks, and the likes. These systems are thus targets of cyber-criminal elements who seek to compromise the integrity, confidentiality, availability or accountability of the IoHT systems. It is therefore necessary to implement smart proactive defense mechanisms against such menaces and promote overall data security. Anomaly based intrusion detection systems leveraging machine learning techniques stand out in this regard. This paper aims to compare the performances of selected machine learning classifier models for detecting attacks against IoHT systems. The ECU-IoHT dataset was adopted for training and testing classifiers such as CatBoost, Extreme Gradient Boosting (XGB) and Light Gradient Boosting Machine (LGBM). Data processing techniques such as the Hold-Out method and Stratified K-fold Cross Validation were employed. Results show that the classifiers were excellent at detecting breaches on IoHT systems. However, for the Stratified K-fold approach, XGB presented the best performance on F1-score and accuracy metrics with 98 percent and 96 percent respectively.
dc.identifier.citationOsa E., Iyekekpolo E. N., Ibhaze A. E., Sadoh W. E. & Orukpe P. E. (2024). “Implementing Stratified K-Fold Technique for Data Intrusion Detection Models in Internet of Health Things Systems". In Proceedings of the International Conference on Artificial Intelligence and Robotics (MIRG-ICAIR 2024), pp. 225-234, MIRG
dc.identifier.isbn978-978-771-680-9
dc.identifier.urihttps://ir.unilag.edu.ng/handle/123456789/13065
dc.language.isoen
dc.publisherMIRG
dc.relation.ispartofseriesMIRG-ICAIR 2024
dc.titleImplementing Stratified K-Fold Technique for Data Intrusion Detection Models in Internet of Health Things Systems
dc.typeArticle
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Implementing Stratified K-Fold Technique for Data Intrusion Detection Models in Internet of Health Things Systems.pdf
Size:
880.42 KB
Format:
Adobe Portable Document Format
Description:
Download
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
1.71 KB
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections
2024 Proceedings