A Data-Driven Approach for Early Detection and Prevention of Diabetes Mellitus Using Machine Learning

Diabetes is a common health issue in India that impacts the quality of life and overburdens healthcare resources. Prevention and early detection are the keys to controlling its effects. This research applies Indian demographic and health statistics to forecast diabetes risk through machine learning. KNearest Neighbors (KNN), Logistic Regression, Decision Tree, Support Vector Machine (SVM), and Random Forest were experimented upon for predictive efficiency. Great preprocessing of data was done to manage missing values and features normalization to make the model reliable. Random Forest yielded the maximum accuracy of 94.50% among the algorithms tested, proving the efficiency in predicting diabetes. The results unveil the promise of machine learning in medicine for early diagnosis and focused treatment approaches.