Prediksi Risiko Stunting Pada Balita Menggunakan Algoritma Decision Tree Berbasis Data Terbuka

Abstract

Stunting remains a major public health challenge with long-term consequences for children’s physical growth and cognitive development. This study aims to develop a classification model for predicting stunting risk in children under five using a Decision Tree algorithm based on simple anthropometric data. The dataset consists of 5,000 training records and 50 testing records, containing variables such as age, sex, and height, along with four nutritional status categories: normal, tall, stunted, and severely stunted. The analysis process includes data preprocessing, splitting of training and testing sets, model training using RapidMiner, and performance evaluation through a confusion matrix as well as accuracy, precision, recall, and f1-score metrics. The results show that the Decision Tree model successfully classified all testing samples correctly. The model achieved 100% accuracy, with precision, recall, and f1-scores reaching maximum values across all classes. These findings demonstrate that the decision rules generated by the model are capable of distinguishing each nutritional status category based on available anthropometric features. Overall, the study indicates that the proposed model has strong potential for use as an early screening tool for stunting risk, particularly in primary healthcare settings that require fast, simple, and data-driven assessment methods.