Predicting stunting status among under-5 children in Rwanda using neural network model: Evidence from 2020 Rwanda...

Area of Science:

• Public Health
• Machine Learning
• Pediatrics

Background:

• Childhood stunting is a significant public health issue in Rwanda, impacting 33.3% of children under five in 2020.
• While machine learning is used for stunting prediction, Artificial Neural Networks (ANNs) are underutilized despite their predictive capabilities.

Purpose of the Study:

• To predict childhood stunting in Rwanda utilizing Artificial Neural Networks (ANNs).
• To leverage the most recent 2020 Demographic and Health Survey (DHS) data for stunting prediction.

Main Methods:

• A multilayer perceptron (MLP) model was developed using the 2020 DHS dataset.
• The dataset was divided into 80% for training and 20% for testing and validation.
• Model performance was evaluated using accuracy, precision, recall, and AUC-ROC, with feature importance analysis.

Main Results:

• The ANN model achieved 72.0% accuracy and an AUC-ROC of 0.84, indicating good predictive performance.
• Maternal height was identified as a critical factor, with shorter mothers associated with increased stunting risk.
• Early initiation of breastfeeding was a significant protective factor, reducing stunting risk.

Conclusions:

• ANNs are a valuable tool for predicting stunting and identifying key risk factors in Rwanda.
• Findings can guide targeted interventions, such as nutritional support and education on nutrition and hygiene.
• The study provides insights applicable to reducing stunting in low- and middle-income countries.