Feature Selection Optimization Using Genetic Algorithm for Naive Bayes-Based Diabetes Mellitus Classification

Sari

Diabetes mellitus is a chronic disease with a steadily increasing prevalence each year and poses the risk of severe complications if not addressed early. Therefore, early detection of diabetes risk plays a vital role in prevention efforts. This study aims to enhance feature selection optimization through the use of a genetic algorithm in the classification of diabetes mellitus patients based on the Naive Bayes method. The genetic algorithm was applied to identify the most significant clinical features from patient data, with the expectation of improving the classification model’s accuracy and efficiency. A dataset comprising 1,557 patient records with 29 initial clinical attributes was utilized. Following preparation and selection stages, 7 key features were chosen for model training. Model performance was evaluated using metrics such as accuracy, precision, recall, and F1-score. The results indicated that the model with selected features achieved an accuracy of 80.99%, precision of 80.99%, recall of 100%, and an F1-score of 89.5%. These findings confirm that genetic algorithms are effective in improving Naive Bayes classification performance for diabetes risk identification. This study is expected to serve as a foundation for the development of more accurate and efficient disease risk prediction systems in the future.