Related Experiment Video
Updated: Sep 7, 2026
Visualizing Efficacy of Pesticides Against Disease Vector Mosquitoes in the Field
Published on: March 16, 2019
Predicting Insecticide-Treated Net Use Among Under-Five Children in Tanzania Using Machine Learning: Evidence From
William Nkenguye1,2,3, Edwin Joseph Shewiyo1,2,3
1Department of Epidemiology and Biostatistics KCMC University Moshi Tanzania.
Background:
Malaria remains a leading cause of morbidity and mortality in sub-Saharan Africa, disproportionately affecting children under five. In Tanzania, where malaria accounts for a significant share of pediatric deaths, the use of insecticide-treated nets (ITNs) is a cornerstone of prevention. However, despite widespread ITN distribution, usage remains suboptimal due to socioeconomic and behavioral factors. This study applied machine learning (ML) methods to predict ITN usage among under-five children in Tanzania using nationally representative survey data.
Methods:
We utilized data from the 2022 Tanzania Demographic and Health Survey (TDHS), comprising 8,319 women with complete information relevant to ITN use. Six supervised ML models-Random Forest, Bagging, Support Vector Machine (SVM), Gradient Boosting Machine (GBM), Naïve Bayes, and Logistic Regression-were developed to predict ITN usage. The dataset was split into training (70%) and testing (30%) sets. Class imbalance was addressed using Synthetic Minority Over-sampling Technique, and model performance was evaluated using AUC, accuracy, sensitivity, specificity, PPV, and NPV. Feature importance was assessed using the Mean Decrease in Gini index.
Results:
The Random Forest model achieved the highest performance (AUC = 0.90, accuracy = 87%, sensitivity = 90%, specificity = 75%), followed closely by Bagging and SVM models. Key predictors of ITN use included wealth index, education level, region, and mosquito net ownership. Socioeconomic and geographic disparities were the strongest contributors to variations in ITN utilization, while pregnancy status and household size were less influential.
Conclusion:
Machine learning offers a powerful approach for identifying determinants of ITN use and targeting high-risk populations. This study demonstrates the potential of predictive modeling to enhance malaria prevention strategies in Tanzania and similar endemic settings. Future work should integrate geospatial and longitudinal data and explore operationalization through digital decision-support systems.
