Classification of nutritional status in toddlers using the support vector machine method

Abstract

Malnutrition among under-fives remains a significant public health issue, especially in Gowa District, South Sulawesi, Indonesia. Accurate determination of the nutritional status of children under five is essential to support targeted health interventions. This study applies the Support Vector Machine (SVM) method to classify the nutritional status of children under five based on 2022 data, considering two categories: malnutrition and good nutrition. This method uses four types of kernels, Linear, Polynomial, Sigmoid, and Radial Basis Function (RBF), to identify non-linear patterns and handle imbalances between data classes. The results show that the RBF kernel performs best, with a classification accuracy of 98.27% and an APER value of 1.73%. This confirms the SVM's ability to handle complex data without assuming linearity, making it a superior approach to other traditional and nonparametric statistical methods. This SVM-based approach offers a significant contribution to the analysis of the nutritional status of individuals under five, not only to improve the accuracy of decision-making in the public health field but also as a basis for further development for the analysis of other health data in the future.