<![CDATA[Type 2 Diabetes Mellitus is a chronic disease that develops gradually and can lead to serious complications—such as heart disease, kidney failure, and blindness—if not detected early. This study aims to evaluate and compare the performance of four machine learning algorithms—Logistic Regression, Random Forest, Multilayer Perceptron, and Deep Neural Network—in predicting the risk of type 2 diabetes based on medical data. The analysis uses the Pima Indians Diabetes dataset, which contains 9.538 patient records and 16 predictor variables. We split the data into training and testing sets using an 80:20 ratio. During training, we performed hyperparameter tuning using Grid Search combined with cross-validation. To evaluate model performance, we applied several metrics, including accuracy, precision, recall, F1-score, Mean Squared Error (MSE), Root Mean Squared Error (RMSE), R², and an analysis of overfitting. The results indicate that the Random Forest model outperformed the others, achieving 100% accuracy, zero classification errors, near-zero prediction error values, and no signs of overfitting. Logistic Regression also performed well, though slightly below the Random Forest. In contrast, the Multilayer Perceptron and Deep Neural Network models showed mild overfitting and higher false negative rates. Based on these findings, we recommend the Random Forest model as the most reliable option for early prediction systems in type 2 diabetes mellitus.]]>
Copyrights © 2025
