Predictive Analysis of Diabetes Using Machine Learning Algorithms

Abstract

Diabetes is a very harmful disease that causes high blood sugar levels and occurs when the blood glucose level is high. Diabetes causes numerous diseases in humans: congestive heart failure, stroke, kidney and eye problems, dental issues, nerve damage, and foot problems. With the recent development in the machine learning concept, it is easy to analyze and predict whether a person is diabetic or not. This research mainly focuses on using several prediction algorithms of machine learning. The algorithms used in this research are k-nearest neighbor, logistic regression, SVM (support vector machine), Gaussian naive Bayes, decision tree, multilayer perceptron, random forest, XGBoost, and AdaBoost. Among these algorithms, the XGBoost performed better than the other algorithms achieving an accuracy of 90%, and the f1 score and Jaccard score were 91% and 86%, respectively. The primary goal of this research is to apply numerous machine learning algorithms to diabetic datasets, analyze their results, and select the best one that performs well.