Improving diabetes mellitus disease patients classification using XGBoost method

Diabetes mellitus is a disease characterized by elevated blood glucose levels due to the inability of the pancreas to produce sufficient insulin, preventing glucose uptake into cells for energy. The increasing prevalence of diabetes globally makes it one of the most pressing public health problems of this century. Its far-reaching impacts—both medically and socially—causes a major burden on health systems and the economy, and risks reducing the quality of life of sufferers. Therefore, serious efforts are needed in developing early detection methods that are accurate, efficient, and widely applicable. Early detection is essential to reduce health risks and overcome potential stigma that can affect access to health services, life insurance, and driving eligibility. This study develops an early detection model for diabetes mellitus using Extreme Gradient Boosting (XGBoost), which was chosen for its ability to handle structured data, reduce overfitting, and provide high predictive accuracy. This method is well suited for this research dataset because it efficiently processes numeric and categorical features while maintaining robustness to noise and missing values. The proposed model achieved an accuracy of 97.12% on the test data, indicating that XGBoost is a promising approach for early diabetes detection and can be effectively applied in clinical settings to support timely diagnosis and treatment planning.