Peer-reviewed article
Exploring Risk Factors and Predictive Modeling of Child Malnutrition in Pakistan Using Machine Learning.
Muhammad Usman Saleem, Muhammad Usman Aslam, Abdul Ghani Khatir, Quanbao Jiang ·
Summary
Written by AI, not yet AICS's view · claude-haiku-4-5-20251001This study used machine learning to identify risk factors for child malnutrition (stunting, wasting, and underweight) in Pakistani children under five, analysing data from the 2017-2018 Demographic and Health Survey. Key risk factors included consanguineous marriages, low wealth status, low maternal education, and geographic location in Sindh and Baluchistan. Random Forest models performed best at predicting malnutrition, while Support Vector Machines showed higher sensitivity for detecting wasted and underweight children. The findings suggest that targeted interventions in high-risk regions should focus on maternal education, water and sanitation access, and poverty reduction.
Key points
- Risk factors for malnutrition included consanguineous marriages, lower wealth status, low maternal education, and geographic disparities in Sindh and Baluchistan
- Random Forest model achieved highest overall accuracy and specificity for predicting all malnutrition indicators
- Support Vector Machine showed higher sensitivity for detecting wasted and underweight children
- Study emphasizes the complex interplay of socio-demographic factors in child malnutrition
- Recommendations include targeted interventions on maternal education, WASH access, and poverty alleviation in high-risk regions
Limitations
This is a cross-sectional analysis based on secondary survey data, so it identifies associations rather than causal relationships and cannot determine temporal sequence or mechanisms. The study does not evaluate any intervention or test outcomes, only describes predictive patterns in existing data.
Abstract
OBJECTIVE: This study aims to identify risk factors and develop predictive models of child malnutrition (stunting, wasting, and underweight) in Pakistani children under five using machine learning approaches. STUDY DESIGN: This cross-sectional design utilized data from the Pakistan Demographic and Health Survey 2017-2018 (PDHS). METHODS: Logistic regression was employed to identify significant socio-demographic and health-related risk factors. Four machine learning models-Random Forest (RF), Support Vector Machine (SVM), Naïve Bayes (NB), and AdaBoost-were applied to predict malnutrition indicators, with performance evaluated based on accuracy, sensitivity, specificity, and F-measure using a two-stage validation strategy (10-fold cross-validation and 80:20 train-test split). RESULTS: Key risk factors identified included consanguineous marriages, lower wealth status, low maternal education, and geographic disparities in Sindh and Baluchistan. Among the machine learning models, Random Forest demonstrated the highest overall accuracy and specificity across all indicators, while SVM showed higher sensitivity for wasted and underweight children. CONCLUSION: The study highlights the complex interplay of socio-demographic factors in child malnutrition and the potential of machine learning models to effectively predict these conditions, underscoring the need for targeted interventions focusing on maternal education, access to clean water and sanitation, and poverty alleviation, particularly in high-risk regions.