The Nutritional Classification of Pregnant Women Using Support Vector Machine (SVM)

Abstract

Determining the nutritional status of pregnant women is one of the efforts to control the condition of pregnant women so that they can adjust their health conditions properly. The health condition of pregnant women can affect the condition of the baby who will be born. This study aims to apply the SVM method to a web-based application to classify the nutritional status of pregnant women based on data obtained from several health centers in the city of Lhokseumawe. SVM functions as the core of the application in charge of classifying the nutritional status of pregnant women based on several features including: age, weight, height, lila, hemoglobin and BMI. While the data class consists of 3 categories, namely: undernourished, normal nutrition and normal nutrition + overweight. Primary data obtained from the field amounted to 355 data which were then divided into two parts with a ratio of 70% training data and 30% testing data. Based on the research conducted, it was found that the application of different kernels in the Support Vector Machine (SVM) will have a different performance impact in classifying data. In this study, the linear kernel has the best performance with an accuracy value of 0.84, the RBF kernel has an accuracy value of 0.83, the polynomial kernel has an accuracy value of 0.72, and the sigmoid kernel has the worst performance with an accuracy value of 0.58