A Case Study on Stunting Classification in Toddlers from Karyajaya South Sumatra Using Extreme Gradient Boosting and K-Nearest Neighbors

Stunting, a consequence of persistent malnutrition, severely hinders the physical and cognitive development of children under five years of age. In Indonesia, the incidence of stunting was 21.6% in 2022, presenting a significant public health challenge. This research seeks to create efficient machine learning (ML) models for identifying stunting status in chil-dren, utilizing a dataset of 3,000 physical measurements from the Karyajaya Health Center in South Sumatra. The study utilizes two classification algorithms: K-Nearest Neighbors (KNN) and Extreme Gradient Boosting (XGBoost). Data preprocessing encompassed feature selection, outlier elimination via the Interquartile Range (IQR) approach, and rectification of class imbalance through the Synthetic Minority Over-sampling Technique (SMOTE) and parameter optimization. The KNN model was set with five neighbors and distance-based weighting, resulting in an accuracy of 99.43% and an F1-score of 93.89%. The XGBoost model, optimized with a calibrated scale-pos-weight parameter, surpassed KNN by achieving an accuracy of 99.93% and an F1-score of 99.62%. These findings illustrate XGBoost's exceptional capacity to manage intricate, unbalanced datasets and underscore its promise as a reliable instrument for early stunting identification. The application of XGBoost can assist policymakers and healthcare professionals in delivering prompt interventions, therefore aiding the national objective of decreasing stunting prevalence to 14% by 2024. Subsequent study could improve model accuracy by integrating supplementary genetic and environmental variables across various demographic contexts.