<![CDATA[Diabetes mellitus is a long-term condition marked by elevated blood sugar levels, which can lead to serious complications such as heart disease, kidney failure, and vision impairment. Early detection plays a vital role in minimizing these risks and enhancing patients' quality of life. This research focuses on assessing the performance of three machine learning algorithms—Random Forest, XGBoost, and LightGBM—in predicting diabetes risk. The dataset utilized originates from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), comprising 768 samples with 9 key features. The research methodology involves multiple stages, including data collection, preprocessing, addressing data imbalance using SMOTE, data splitting for training and testing, algorithm implementation, and model evaluation through accuracy, precision, recall, F1-score, and Area Under the Curve (AUC) metrics. Findings reveal that Random Forest delivers the highest performance with an AUC score of 86%, followed by XGBoost (83%) and LightGBM (82%). With its strong accuracy, this model holds potential as a valuable tool for early diabetes diagnosis, contributing to faster and more precise medical decision-making.]]>
Copyrights © 2025
