Machine Learning Based Prediction of Edematous Malnutrition in Afghan Children

Abstract

Proper nutrition is one of the necessities of human health. Malnutrition in children is an ecumenical problem that causes mortality and morbidity in children. Afghanistan is one of the countries which suffers from child severe acute malnutrition. In this study, a machine learning based model was proposed to predict the severity of edematous malnutrition in children between the age of 1–59 months in the context of Afghanistan. Random Forest, J48, and Naïve Bayes classifiers were applied to the malnutrition-related data, which was collected from two hospitals in Afghanistan. The Random Forest technique obtained outstanding results with the highest accuracy of 97.14% and the performance of J48 was also moderate with 94.51% accuracy. This study explores how machine learning classification techniques can classify the edematous malnutrition in under 5 (U5) children. Overall, the findings of the proposed method demonstrate that our model with a robust result provides a potential mechanism for prediction of nutritional oedema using machine learning classification algorithms for Afghan U5 children that helps policy makers for child’s malnutrition.