A Machine Learning based Approach to Detect Early Stage Diabetes Prediction

Diabetes is one of the nation's primary causes of the spike in mortality rates. The surge in diabetes has been directly associated with an unhealthy lifestyle, urbanization, obesity/overweight, genetics, hormonal imbalance, poor diet, smoking, and alcoholism. Diabetes is very much harmful if left unidentified over the long term, which can lead to life- threatening difficulties like stroke and heart diseases. Through the application of Machine Learning algorithms to real-life problems, it is possible to come up with efficient, effective, and tailor-made solutions to detect diabetes at early stages. In this research paper, several ML models are compared and analyzed for early diabetes detection. The various categorization techniques used for our model development are SVM, DT, Random Forest, XGBoost, KNN, Logistic Regression. Through grid search, the hyperparameters of the models are tuned to achieve optimal performance. The proposed algorithm's performance is evaluated using various performance metrics like precision, Accuracy, Recall and F1-Score and ROC-AUC Curve.