Peer-reviewed article
Predicting stunting status among under-5 children in Rwanda using neural network model: Evidence from 2020 Rwanda demographic and health survey.
Similien Ndagijimana, Ignace Kabano, Emmanuel Masabo, Jean Marie Ntaganda ·
Summary
Written by AI, not yet AICS's view · claude-haiku-4-5-20251001This study used machine learning to predict stunting (inadequate growth) among children under five in Rwanda using data from the 2020 Demographic and Health Survey. The researchers trained an artificial neural network model that achieved 72% accuracy in identifying which children were at risk of stunting. The analysis found that maternal height and early breastfeeding initiation were among the strongest predictors of child stunting status. The study suggests this predictive approach could help identify vulnerable children and target nutrition programs to mothers and families.
Key points
- The neural network model achieved 72% accuracy with an AUC-ROC of 0.84 in predicting stunting status
- Maternal height was identified as the most important negative predictor: lower maternal height associated with increased child stunting risk
- Early breastfeeding initiation was identified as a key protective factor reducing stunting risk
- Stunting affected approximately 33.3% of children under five in Rwanda in 2020
- The study suggests machine learning could support targeted nutritional interventions for pregnant and lactating mothers
Limitations
This is a predictive modeling study, not an intervention evaluation, so it cannot demonstrate whether any particular program actually reduces stunting. The study identifies associations between factors and stunting status but does not establish causation or test the effectiveness of recommended interventions.
Abstract
BACKGROUND: Stunting is a serious public health concern in Rwanda, affecting around 33.3% of children under five in 2020. The researchers have employed machine learning algorithms to predict stunting in Rwanda; however, few studies used ANNs, despite their strong capacity to predict stunting. The purpose of this study was to predict stunting in Rwanda using ANNs and the most recent DHS data from 2020. METHODS: DHS 2020 dataset was used to train and test an ANN model for predicting stunting in children. The dataset, which included various child, parental, and socio-demographic characteristics, was split into 80% training data and 20% testing and validation data. The model utilised a multilayer perceptron (MLP). Model performance was assessed using accuracy, precision, recall, and AUC-ROC. Feature importances were determined and highlighted the most critical predictors of stunting. RESULTS: An overall accuracy of 72.0% on the test set was observed, with an AUC-ROC of 0.84, indicating the model's good performance. Factors appear to contribute to stunting among the negative value aspects. First and foremost, the mother's height is important, as a lower height suggests an increased risk of stunting in children. Positive value characteristics, on the other hand, emphasise elements that reduce the likelihood of stunting. The timing of the initiation of breastfeeding stands out as a crucial factor, showing that early breastfeeding initiation has been linked with a decreased risk of stunting. CONCLUSIONS: These findings suggest that ANNs can be a useful tool for predicting stunting in Rwanda and identifying the most important associated factors for stunting. These insights can inform targeted interventions to reduce the burden of stunting in Rwanda and other low- and middle-income countries. Potential targeted interventions include nutritional support programs for pregnant and lactating mothers, and providing educational programs for parents on nutrition and hygiene.