A Supervised Learning Approach by Machine Learning Algorithms to Predict Diabetes Mellitus (DM) Risk Score

Abstract

AbstractThe utilization of artificial intelligence (AI) has become a valuable part of medical research. Diabetes is one of the top maladies on the planet, and it is a long-term disease that happens if the body cannot utilize glucose appropriately. This paper indicates the demand for the right procedures, advancement, and usage of successful well-being observing strategies ought to be selected to battle against diabetes. So early discovery and treatment with the utilization of different procedures must be preferred. Our model consists of four machine learning algorithms which are K-Nearest Neighbor, Random Forest, Decision Tree, and Logistic Regression. These algorithms are used on a dataset of 15,000 diabetes mellitus patients along with nine features, where the results are compared to show which algorithm gives the best accuracy, including statistical analysis to get a more perfect result of our model. Among all four algorithms, the random forest gives the best accuracy of around 92%, where the rest of the algorithms give between 78 and 90%. Hopefully, this study could be very helpful in medical science to predict the risk score of diabetes mellitus (DM) early and to classify them. KeywordsDiabetes Mellitus (DM)Linear Discriminant Analysis (LDA)Logistic RegressionRandom ForestDecision TreeKNNData analysisPredictionROC-AUC curveStatistical analysis