Using Machine Learning Techniques to Predict Type 2 Diabetes

Abstract

Diabetic patients face multiple risks of potential complications;
therefore, early diagnosis is so important to avoid these consequences. In
healthcare scientific research, the literature revealed that machine learning
techniques are widely used to diagnose many diseases, including diabetes.
This paper aims to predict diabetes using some machine-learning techniques.
A dataset from Kaggle was used in the study, which is originally from the
National Institute of Diabetes and Digestive and Kidney Diseases. Random
Forest, Logistic Regression, and Support Vector Machine techniques were
used for classification. The results revealed that the highest obtained
accuracy value was with the 0.831 algorithm Logistic Regression and
Random Forest.