DIABETES MELLITUS PREDICTION USING SUPERVISED MACHINE LEARNING TECHNIQUES
Abstract
High blood glucose levels are the cause of diabetes mellitus, a chronic and potentially fatal illness. It might result in serious side effects such as kidney failure, neuropathy, and cardiovascular disorders if undetected or mistreated. In order to reduce these dangers, early identification is essential. Using five supervised algorithms—Logistic Regression, Decision Tree, Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Random Forest—this work investigates the creation of a machine learning system for early diabetes prediction. Logistic Regression obtained the maximum accuracy of 96% in experiments on the Bangladesh Diabetes Dataset (BDD). A web-based interface was developed using Flask to implement the model, offering a userfriendly platform for managing and predicting diabetes.
Keywords
Explainable Machine Learning (LIME) for Di- abetes, Logistic Regression Dominance in Health Analytics