Successive Suppression of Viral Load in ART-Treated Patients: Machine Learning Algorithms and Dimensionality Techniques Prediction

The use of Antiretroviral (ART) drugs significantly reduces Viral Load (VL) when adhered to properly. However, non-compliance can result in progression to Acquired Immune Deficiency Syndrome (AIDS) and potentially death. In this study, a combination of Machine Learning (ML) algorithms, Traditional Regression (TR), and Exploratory Data Analysis (EDA) was utilized for data modeling. The ML models included Multilayer Perceptron (MLP), Support Vector Machine (SVM), and Adaptive Neuro-Fuzzy Inference System (ANFIS). Traditional regression models consisted of Multiple Linear Regression (MLR) for predictive purposes and Logistic Regression (LR) for classification tasks. Multiple Correspondence Analysis (MCA) of EDA was employed to predict distance, covering model accuracy. The study utilized simple random sampling to select 1500 HIV patients on ART at Federal Teaching Hospital, Gombe State, Nigeria (FTHG). The MLP model demonstrated superior predictive performance in the calibration phase with an R-square (R2) of 0.920209 and Mean Squared Error (MSE) of 0.002150. In the verification phase, MLR achieved a higher R2 of 0.913961 and a lower MSE of 0.002234 than other models. The LR regression classifier model explained 73.3% of the variance in predicting patients’ commitment to ART medication. Significant predictors included age, viral load, gender, marital status, age group, and drug type (p < 0.05). However, treatment compliance and defaulters demonstrated a sensitivity of 85.64%, specificity of 93.47%, positive predictive value of 87.92%, negative predictive value of 91.97%, and overall accuracy of 90.53%. The ROC curve indicated that the viral load score does not significantly predict outcomes (p > 0.05). Similarly, the MCA indicated an inertia of 45.5%, suggesting that the dimensions significantly contribute to total variance, with an eigenvalue of 5.47, and Cronbach's alpha was 69.1%, reflecting acceptable internal consistency.

Authors

Institutions

Publication Details

Journal
Journal of Progress in Artificial Intelligence
Published
2026-11-12
DOI
https://doi.org/10.68152/2526
Primary Topic
HIV/AIDS Research and Interventions
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Successive Suppression of Viral Load in ART-Treated Patients: Machine Learning Algorithms and Dimensionality Techniques Prediction

Isah Muazu, İlker Etikan, Kabiru Bala, Ibrahim Danladi Yakoko
Journal of Progress in Artificial Intelligence
HIV/AIDS Research and Interventions
article

Successive Suppression of Viral Load in ART-Treated Patients: Machine Learning Algorithms and Dimensionality Techniques Prediction

Isah Muazu, İlker Etikan, Kabiru Bala, Ibrahim Danladi Yakoko
article en

Abstract

The use of Antiretroviral (ART) drugs significantly reduces Viral Load (VL) when adhered to properly. However, non-compliance can result in progression to Acquired Immune Deficiency Syndrome (AIDS) and potentially death. In this study, a combination of Machine Learning (ML) algorithms, Traditional Regression (TR), and Exploratory Data Analysis (EDA) was utilized for data modeling. The ML models included Multilayer Perceptron (MLP), Support Vector Machine (SVM), and Adaptive Neuro-Fuzzy Inference System (ANFIS). Traditional regression models consisted of Multiple Linear Regression (MLR) for predictive purposes and Logistic Regression (LR) for classification tasks. Multiple Correspondence Analysis (MCA) of EDA was employed to predict distance, covering model accuracy. The study utilized simple random sampling to select 1500 HIV patients on ART at Federal Teaching Hospital, Gombe State, Nigeria (FTHG). The MLP model demonstrated superior predictive performance in the calibration phase with an R-square (R2) of 0.920209 and Mean Squared Error (MSE) of 0.002150. In the verification phase, MLR achieved a higher R2 of 0.913961 and a lower MSE of 0.002234 than other models. The LR regression classifier model explained 73.3% of the variance in predicting patients’ commitment to ART medication. Significant predictors included age, viral load, gender, marital status, age group, and drug type (p < 0.05). However, treatment compliance and defaulters demonstrated a sensitivity of 85.64%, specificity of 93.47%, positive predictive value of 87.92%, negative predictive value of 91.97%, and overall accuracy of 90.53%. The ROC curve indicated that the viral load score does not significantly predict outcomes (p > 0.05). Similarly, the MCA indicated an inertia of 45.5%, suggesting that the dimensions significantly contribute to total variance, with an eigenvalue of 5.47, and Cronbach's alpha was 69.1%, reflecting acceptable internal consistency.

Journal of Progress in Artificial IntelligenceVol. 02(04)
Taraba State University (NG), Near East University (CY)
Good health and well-being
HIV/AIDS Research and Interventions
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon