Application of XGB Classifier for Obesity Rate Prediction

Abstract

According to the Ministry of Health, the percentage of the population in Indonesia who are overweight is 13.5% for adults aged 18 years and over, while 28.7% are obese with BMI>=25 and obese with BMI>=27 as much as 15.4%. Meanwhile, at the age of children 5-12 years, 18.8% were overweight and 10.8% were obese. From these data, early detection of obesity levels is needed. From these data, prevention is needed so that the percentage of the population who experience obsediness can decrease, one of the efforts that can be done is to do early detection of obesity, to do early detection of obesity can be done using Machine Learning. In this study, it was discussed about the prediction of obestias levels using 7 (seven) models, namely Naive Bayes (NB), Random Forest (RF), K-NN, Decision Tree Classifier (DTC), SVM, XGB Classifier (XGB), Logistic Regression (LR) from the seven models used to predict the obesity level of XGB Classifier (XGB) which has the highest accuracy, namely Accurasy 0.96, with an f1-score of 0.96, Precission and recall 0.96.

Keywords

• obesity
• xgb classifier
• crips-dm