Optimized Prediction Model for Type 2 Diabetes Mellitus Using Gradient Boosting Algorithm

Diabetes is a growing disease which has approximately affected 381 million people over the globe. Among diabetic patients, 30% of the adult population is diagnosed with Type 2 Diabetes Mellitus (T2DM). Predicting T2DM at early stages can help avoid serious complications at later stages. Data mining (DM) and machine learning (ML) based techniques are gaining popularity in disease diagnosis, reasons discovery, and decision making. Various research studies using DM are available in the literature to predict diabetes but there is room to improve the performance. In this study, a gradient boosting algorithm is proposed for the prediction of T2DM. We have used the PIMA Indian Diabetes dataset collected by an American institute UCI for analyzing our prediction model with 10-fold cross-validation. Our proposed model exhibits better performance as compared to various other models proposed in the literature. The proposed model can support medical experts in the decision-making and prediction of T2DM.