Related Experiment Video
Updated: Jul 3, 2026
A Point-of-Care Method with Integrated Decision Support Tool to Estimate Anemia at Population Level
Published on: January 19, 2024
Improved prediction of childhood anemia using hybrid ensemble learning and dual-level explainability
Ashima Kukkar1, Julius Moinget Loibor2, Mehmet Akif Cifci3,4
1Institute of Engineering and Technology, Chitkara University, Punjab, India.
Insights
Early childhood anemia prediction is crucial for public health. An interpretable machine learning framework accurately detects anemia in children, aiding clinical decisions in underserved regions.
Area of Science:
- Machine Learning in Public Health
- Computational Epidemiology
- Pediatric Health Informatics
Background:
- Millions of children under five, especially in low- and middle-income countries, suffer from preventable anemia.
- Early anemia detection is critical for improving public health outcomes and preventing long-term developmental issues.
Purpose of the Study:
- To propose an interpretable machine learning framework for the early prediction of childhood anemia.
- To utilize structured healthcare data for anemia risk assessment in pediatric populations.
Main Methods:
- A stacked ensemble architecture integrating TabNet, XGBoost, and Multi-Layer Perceptron (MLP) with logistic regression as a meta-learner.
- Hyperparameter optimization using GridSearchCV and Optuna, with feature importance analysis via SHAP and individual prediction explanations using LIME.
- Training and testing on the Tanzania Demographic and Health Survey (DHS) dataset, with external validation using NFHS (India) data.
Main Results:
- The ensemble model achieved 98.5% accuracy, outperforming individual models with high precision, recall, and F1-scores.
- GridSearchCV and Optuna demonstrated optimal performance for hyperparameter tuning. Key predictors identified include child age, wealth index, and breastfeeding status.
- External validation confirmed cross-regional generalizability, showing close agreement with observed anemia prevalence trends.
Conclusions:
- The developed framework is accurate, transparent, and adaptable for early anemia detection.
- It can significantly assist clinical decision-making, particularly in resource-limited settings.
- The tool offers valuable support for global public health initiatives aimed at combating childhood anemia.
Purpose:
Millions of children under five, particularly in low- and middle-income countries, suffer from preventable anemia, making early detection critical for improving public health outcomes. This study proposes an interpretable machine learning framework for early prediction of childhood anemia using structured healthcare data.
Methods:
The proposed approach integrates TabNet, XGBoost, and Multi-Layer Perceptron (MLP) within a stacked ensemble architecture, with logistic regression as a meta-learner. Hyperparameter optimization is performed using GridSearchCV and compared with RandomizedSearchCV, Optuna, Genetic Algorithms (GA), Particle Swarm Optimization (PSO), and Ant Colony Optimization (ACO). SHAP provided global feature importance, and LIME explained individual predictions. The system was trained and tested using the Tanzania Demographic and Health Survey (DHS) dataset.
Results:
Experimental results on the Tanzania Demographic and Health Survey (TDHS 2022) dataset demonstrate that the proposed ensemble significantly outperforms individual models, achieving an accuracy of 98.5% and high precision, recall, and F1-scores across both classes. Among optimization strategies, GridSearchCV and Optuna provide the most consistent and optimal performance. SHAP-based global analysis identifies key predictors such as child age, wealth index, and breastfeeding status, while LIME offers instance-level explanations, enhancing model transparency and clinical interpretability. Comparative evaluation with baseline models and ablation analysis confirms the effectiveness of the stacked architecture and meta-learning strategy. External validation using NFHS (India) data shows close agreement (±1-2%) with observed anemia prevalence trends, indicating cross-regional generalizability. The study also discusses ethical considerations, fairness implications, and deployment strategies for integration into healthcare systems.
Conclusions:
The proposed framework is accurate, transparent, and adaptable. It supports early anemia detection and can assist clinical decision-making, especially in under-resourced regions, making it a valuable tool for global public health applications.