Under-Five Children's Nutritional Status Prediction Using Naïve Bayes and Decision Tree Based on Anthropometric Data and Mother–Child Class Participation

Abstract

Nutritional status is an important indicator of the health and development of children under five, making early identification essential for supporting appropriate nutritional interventions. This study aimed to develop and compare the performance of the Naïve Bayes and Decision Tree algorithms in classifying the nutritional status of children under five based on anthropometric measurements and participation in the mother and Children Under Five Class program in Mlonggo District, Indonesia. A quantitative approach was applied using the Cross-Industry Standard Process for Data Mining (CRISP-DM) framework. The initial dataset contained 4,588 records, of which 4,563 valid records remained after the preprocessing stage. Model performance was evaluated using accuracy, precision, recall, F1-score, confusion matrix, ROC curve, and McNemar’s test. The results showed that the Decision Tree algorithm achieved a higher cross-validation accuracy of 90.90% compared with 87.59% for Naïve Bayes. The testing results also demonstrated that Decision Tree consistently outperformed Naïve Bayes across the evaluation metrics. Therefore, Decision Tree was selected as the most suitable model for nutritional status classification. The model was subsequently implemented in a web-based application supporting individual and batch prediction, along with the presentation of prediction results and recommended health interventions. The system can support healthcare workers in conducting nutritional status assessments more efficiently and objectively.