<![CDATA[Early childhood represents a crucial stage in human growth and development, making nutritional monitoring essential to prevent long-term health issues such as stunting. This research focuses on developing a nutritional status classification model for toddlers by applying the Decision Tree algorithm with four categories: severely stunted, stunted, normal, and tall. The dataset, obtained from Kaggle, contains 121,000 records and includes attributes such as age, gender, and height. The study was carried out through several phases, starting with data preprocessing to handle missing values, detect outliers, and balance class distribution, followed by model training in R-Studio, and performance evaluation using accuracy, precision, recall, and F1-score. The experimental results demonstrate that the model achieved an accuracy of 89.75%, precision of 89.74%, recall of 89.83%, and an F1-score of 89.78%. The novelty of this study lies in implementing a multi-class classification approach on a large and representative dataset, integrating oversampling and parameter optimization techniques to improve predictive performance, and conducting feature importance analysis that highlights the significant influence of height and age in determining nutritional status. Therefore, this work not only provides a reliable classification model but also contributes practical insights for developing early detection systems to support stunting prevention among toddlers.]]>
Copyrights © 2025
