<![CDATA[Toddlerhood is a critical developmental period that requires precise nutritional monitoring. However, automated classification systems are often challenged by imbalanced data, which makes minority classes difficult to detect accurately. This study aims to optimize a Support Vector Machine (SVM) using a polynomial kernel to improve detection sensitivity for critical classes. By excluding BMI features to avoid redundancy, the proposed model achieved an accuracy of 98%. The main novelty of this research lies in its achievement of an F1 Macro Score of 0.86, confirming that the model provides balanced and reliable classification performance across all nutritional status categories. These results demonstrate the model’s superiority in identifying the minority classes of Severe Malnutrition and Undernutrition more effectively than previous studies. Therefore, the model is highly recommended as an objective decision support system for the early detection of stunting.]]>
Copyrights © 2026
