Prediction of Diabetes Mellitus Disease Using Supervised Machine Learning Models

Diabetes is a noxious disease which might result in a global health care disaster. More than 46.3 crore people around the globe have diabetes, according to the IDF (International Diabetes Federation), which is expected to rise to 64.3 crore till 2030 and 78.3 crore by 2045. Diabetes Mellitus is an illness that occurs due to increased blood glucose levels. Regular urination, an increased desire for water, and an increase in appetite are symptoms of this. Types 1 and 2 are the most prevalent forms of this disease, and there is also a variety that develops during pregnancy and is referred to as gestational diabetes mellitus. Large datasets are analyzed by machine learning algorithms, which utilize statistical models to find trends. Our primary goal is to assess and contrast the various machine learning techniques currently in use in order to determine which approach is the most effective at producing high prediction accuracy for the huge dataset. The diagnosis of diabetes in an individual is made possible by algorithms like Random Forest, DT, NB, SVM, and XGBoost. In order to identify the optimal methodology, efficiency matrices and accuracy comparisons are computed for each. With this approach, XGBoost gives the 96% and Randon forest gives 94% approximate accuracy that making it much superior to other algorithms.