Application of machine learning on the prediction of diabetes screening outcomes using common symptoms in resource-poor settings
Journal Health NPEPS · 12.05.2026 · 11 (1)
DOI: 10.30681/2526101014416 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal Health NPEPS, online erschienen am 12.05.2026. Verfasst von Yashik Singh. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Themen: Künstliche Intelligenz in der Pflege, Diabetes mellitus
Abstract (Originalsprache)
Objective: to evaluate whether a machine learning algorithm could accurately predict diabetes screening outcomes using easily recognized symptoms rather than laboratory or physiological measurements. Method: de-identified patient data from Sylhet Diabetes Hospital in Bangladesh, were used. Fourteen symptoms, along with age and gender, were input into nine supervised machine learning models: Random Tree, C4.5, C-RT, CS-MC4, Linear Discriminant Analysis, Rule Induction, Decision List, ID3, and Partial Least Squares. Results: the models effectively predicted diabetes status based on symptoms, achieving an average accuracy of 94.2% ± 4, TPR of 93.4% ± 4, TNR of 95% ± 5, FPR of 4.6% ± 5, FNR of 6.6% ± 5, and an F-measure of 94.3% ± 4. Conclusion: the Random Tree algorithm performed best and shows strong potential for development into a user-friendly screening tool to encourage timely medical evaluation.
Zitieren
APA
Singh, Y. (2026). Application of machine learning on the prediction of diabetes screening outcomes using common symptoms in resource-poor settings. Journal Health NPEPS, 11(1). https://doi.org/10.30681/2526101014416
Vancouver
Singh Y. Application of machine learning on the prediction of diabetes screening outcomes using common symptoms in resource-poor settings. Journal Health NPEPS. 2026;11(1). doi:10.30681/2526101014416
BibTeX
@article{Singh2026,
author = {Singh, Yashik},
title = {Application of machine learning on the prediction of diabetes screening outcomes using common symptoms in resource-poor settings},
journal = {Journal Health NPEPS},
year = {2026},
volume = {11},
number = {1},
doi = {10.30681/2526101014416}
}
EUVIMED