Multi-Class Classification of Metabolic Syndrome Group Using Gradient Boosting

Abstract

Currently, metabolic syndrome, the leading cause of most of non-communicable diseases, has been overlooked because it is common and does not cause illness at the early stage. The purpose of this study was to classify the group of metabolic syndromes including diabetes, cardiovascular disease, and high blood pressure using machine learning based method. The data was collected from three sub-districts in the province of Chiang Rai: Mae Khao Tom, Nang Lae, and Tha Sud totally 1,605 records. The Gradient Boosting is proposed for the multi-class classification model. From the comparison results with other existing methods, including Random Forest, Extra Trees, K-nearest Neighbor, Support Vector Machines, and Decision Trees, Gradient Boosting outperforms other existing methods having 95.27% accuracy, 95.24% precision, 95.26% recall, and 95.24% F1 score, respectively.