A Predictive Model for Diabetes Diagnosis Using Machine Learning

Diabetes is a chronic metabolic disorder that results in persistently elevated blood glucose levels. It is one of the most prevalent and life-risking illnesses all over the globe today. Its seriousness and risk factors can be greatly reduced with an early and proper prediction. Machine Learning (ML) in the health sector, and especially in disease prediction, has gained immense popularity. The model suggested uses Logistic regression, Naïve Bayes, K-nearest neighbor and Gradient Boosting Classifier for diabetes prediction, and the Gradient Boosting Regressor is utilized to create personalized insulin dosage predictions for diabetic patients. The model performance is tested using ML metrics such as accuracy, precision, recall, and F1-score. Diagnostic data are employed to predict diabetes based on the Pima Indians Diabetes Database (PIDD). The precision is reported to be $95 \%, 89 \%, 92 \%$, and 98 % based on LR, NB, KNN, and GBC algorithms.