<![CDATA[Diabetes Mellitus is a disease that occurs due to disorders of carbohydrate, fat and protein metabolism associated with a lack of performance of insulin secretion. Diabetes is a degenerative disease that requires appropriate and serious treatment efforts. The effects lead to various complications of other serious diseases such as heart disease and stroke. Erectile dysfunction, kidney failure, nervous system damage, etc. Because there are so many impacts caused by diabetes, it is important to study this disease. The benefit of this study is to prevent the occurrence of severe complications and can help medical personnel in predicting this disease early and reduce the cost burden that arises due to this problem. The purpose of this study is to determine the level of accuracy resulting from the use of feature selection with genetic algorithms and nave Bayes. In this study, predictions will be made using hyperparameter tuning with genetic algorithms and Naive Bayes optimization by performing feature selection. After conducting related research, it was found that the accuracy of 17 features using a genetic algorithm was better than modeling with 10 features. By using 17 features and hyperparameter tuning with genetic algorithm and naive Bayes modeling, the accuracy is 93.2%. By using 17 features without feature selection, the accuracy is 91.2%, there is an increase in accuracy of 1.5%.]]>
Copyrights © 2022
