Novita Ranti Muntiari, Khoirun Nisa, Arif Setia Sandi A, Imam Ahmad Ashari, Kharis Hudaiby Hanif, Ramadya Wahyu Dwinanto
Diabetes is a long-term or chronic condition characterized by high or above-normal blood glucose levels. People who combine physical exercise with a high-calorie, high-fat diet that lacks fiber are at risk of developing diabetes. Machine Learning is a type of artificial intelligence that may be used to categorize someone's diabetes. The system will learn to recognize several diabetic factors automatically using the provided medical record data set. The Random Forest, Support Vector Machine, and K-Nearest Neighbor algorithms are compared in this study. In order to prove which is superior, Random Forest is combined with several decision tree possibilities. The Support Vector Machine is used to find a dividing line so that it can distinguish between the two classes of diabetic patients who are treated optimally and those who are not. K-Nearest Neighbor is a classification algorithm that uses characteristics and training samples to categorize new objects. The test results from the comparison of three approaches for categorizing diabetes show that K-Nearest Neighbor (KNN) has the lowest level of accuracy compared to the two methods (89.6%), Support Vector Machine algorithm has 91.5 percent accuracy, and Random Forest algorithm has 93.5 percent accuracy. As a result, the Random Forest algorithm is a suitable and effective approach for assessing the level of accuracy in diabetes diagnosis. © 2023 Author(s).
Harapan Bangsa University, Purwokerto, 53182, Indonesia
Research at a Glance
Register to unlock
Topics & SDG Alignment
Register to unlock
Collaboration
Register to unlock
Author Profile (Selected)
Register to unlock
References Overview
Register to unlock
Journal & Source
Register to unlock
Metadata & Integrity
Register to unlock