Use of machine learning models to predict viral suppression and associated factors
Applied Mathematical Sciences · 2025 · 19 (8) · 359-369
DOI: 10.12988/ams.2025.919280 ↗Kurz gefasst
Fachartikel in der Zeitschrift Applied Mathematical Sciences, 2025. Verfasst von Idah Orowe. Die Arbeit wurde bisher 3-mal zitiert (Crossref).
Abstract (Originalsprache)
Achieving timely viral suppression among individuals undergoing antiretroviral therapy is critical for improving health outcomes and reducing HIV transmission.This study employs machine learning survival models: Support Vector Survival (SVS), Random Survival Forest (RSF), Gradient Boosted Survival (GBS) and Extreme Gradient Boosted Survival (XGBS) to estimate the time to viral suppression and determine key predictive factors among ART recipients. The SVS model adapts support vector machine principles to censored survival data, enabling the modeling of complex, nonlinear relationships in high-dimensional datasets. The RSF model, a nonparametric ensemble approach, constructs multiple survival trees to capture intricate variable interactions without relying on proportional hazard assumptions. The GBS model iteratively enhances survival predictions through gradient boosting, optimizing loss functions tailored to censored data. Model evaluation utilized performance metrics including the concordance index (C-index), integrated Brier score, and time-dependent area under the curve (AUC). Findings demonstrated that ensemble-based models, particularly RSF and GBS, outperformed SVS in predictive accuracy and robustness, effectively identifying key determinants of viral suppression such as baseline viral load, adherence levels, age, and treatment regimen. These results highlight the effectiveness of machine learning survival techniques in improving prediction of treatment outcomes and strengthening evidence-based decision-making in ART program management.
Zitieren
APA
Orowe, I. (2025). Use of machine learning models to predict viral suppression and associated factors. Applied Mathematical Sciences, 19(8), 359-369. https://doi.org/10.12988/ams.2025.919280
Vancouver
Orowe I. Use of machine learning models to predict viral suppression and associated factors. Applied Mathematical Sciences. 2025;19(8):359-369. doi:10.12988/ams.2025.919280
BibTeX
@article{Orowe2025,
author = {Orowe, Idah},
title = {Use of machine learning models to predict viral suppression and associated factors},
journal = {Applied Mathematical Sciences},
year = {2025},
volume = {19},
number = {8},
pages = {359--369},
doi = {10.12988/ams.2025.919280}
}
EUVIMED