Perbandingan Metode Classification and Regression Trees (CART) dengan Naïve Bayes Classification (NBC) dalam Klasifikasi Status Gizi Balita di Kelurahan Pagesangan Barat

Abstract

This study aims to compare the Classification and Regression Trees (CART) and Naïve Bayes Classification (NBC) methods in classifying the nutritional status of toddlers in West Pagesangan by looking at their accuracy and also knowing the variables that influence the classification of toddler nutritional status. The data used in this study were toddlers who come to the posyandu in May 2019, with predictor variables used namely gender, ages, weight, mother’s employment status, mother’s education level, number of children and parents income. The result showed that Naïve Bayes Classification (NBC) is better in classifying the nutritional status of toddlers in West Pagesangan than Classification and Regression Trees (CART). This can be seen from the accuracy values obtained with three comparisons of training data and testing data. In the comparison of 90% of training data: 10% of testing data, obtained an accuracy value of 90% for NBC and 85% for CART, in the comparison of 80% of training data: 20% of testing data, obtained an accuracy value 0f 82.5% for NBC and 80% for CART, while in comparison 70% traing data : 30% testing data, obtained an accuracy value 72% for NBC and 70%for CART. This study also showed that significant variables the classification of nutritional status of toddlers in West Pagesangan village are age, gender, weight and parents income.