Artificial Intelligence Approaches for Predicting Diabetes in Egypt

Abstract

One major public health concern in Egypt is the increasing incidence of diabetes mellitus. It is essential to recognize problems early and treat them effectively [1]. This work applies several machine learning methods to predict diabetes risk using a dataset from Egyptian diabetes and endocrinology clinics. Features including age, BMI, medical history, and other health markers are included in the dataset. Using performance criteria such as confusion matrix, F1-score, recall, accuracy, and precision, we assessed various models including K-Neighbors, Gaussian Naive Bayes, Bernoulli Naive Bayes, Extra Trees, SVC, and Logistic Regression. The findings indicate that diabetes can be accurately predicted using machine learning. Logistic Regression, with a cross-validated accuracy of 0.965, test accuracy of 0.957, precision of 0.94, recall of 0.90, and an F1-score of 0.92, proved to be the most effective model for this dataset.