COMPARISON OF ENSEMBLE LEARNING ALGORITHM IN CLASSIFYING EARLY DIAGNOSTIC OF DIABETES

Authors

  • Okta Jaya Harmaja Program Studi Sistem Informasi, Fakultas Teknologi dan Ilmu Komputer, Universitas Prima Indonesia
  • Irvan Prasetia Program Studi Sistem Informasi, Fakultas Teknologi dan Ilmu Komputer, Universitas Prima Indonesia
  • Yosi Victor Hutagalung Program Studi Sistem Informasi, Fakultas Teknologi dan Ilmu Komputer, Universitas Prima Indonesia
  • Hendra Ardanis Sirait Program Studi Sistem Informasi, Fakultas Teknologi dan Ilmu Komputer, Universitas Prima Indonesia

DOI:

https://doi.org/10.34012/jurnalsisteminformasidanilmukomputer.v7i1.4054

Abstract

Diabetes is a significant public health problem and affects millions of people worldwide. This study will perform a comparative analysis of three ensemble learning algorithms (Random Forest, AdaBoost, and XGBoost) in classifying diabetes diagnoses. Based on the research that has been carried out, it is concluded that the model with the highest accuracy is Random Forest with a value of 0.86, XGBoost with a value of 0.85, and AdaBoost with a value of 0.82. It can also be concluded that the three models perform well and can be used to classify diabetes. Based on the visualization of the results of Feature Importance that has been made, it can be concluded that the Random Forest and XGBoost algorithms have in common the 3 most important features, namely Glucose, BMI and Age. As for AdaBoost, the 3 most important features are DPF, BMI and Glucose.

Downloads

Published

2023-08-24

How to Cite

[1]
O. J. Harmaja, I. Prasetia, Y. V. Hutagalung, and H. A. Sirait, “COMPARISON OF ENSEMBLE LEARNING ALGORITHM IN CLASSIFYING EARLY DIAGNOSTIC OF DIABETES”, JUSIKOM PRIMA, vol. 7, no. 1, pp. 218-231, Aug. 2023.

Most read articles by the same author(s)