Diabetes Prediction using SMOTE and Machine Learning

This research work explores highly sophisticated diabetes prediction algorithms employing the PIMA Indian Diabetes dataset. Proposed research intends to explore the influence of model update, assessment criteria, and data preparation on prediction algorithms. In this extensive research, a pre-selected dataset coupled with feature scaling, stratified selection, and oversampling is employed to tackle the issue of class imbalance. Through the use of advanced machine learning models like Random Forest, the research illustrates how modifying a component's features might enhance estimate accuracy. Using stratified shuffle split validation, the performance of the model is examined and discover large gains in accuracy, F-measure, precision, recall, and AUC. Proposed work underlines the necessity of data preparation for accurate diabetes prognosis and offers an example of outstanding Random Forest model construction.