<![CDATA[Diabetes mellitus is a dangerous disease caused by high sugar levels (hyperglycemia). Hyperglycemia can cause sufferers to experience chronic disease, damage to organs in the body. Diabetes mellitus is a dangerous disease, so it is very interesting to classify diabetes mellitus using the Naïve Bayes method with Backward Elimination (BE) feature selection. The Diabetes mellitus dataset used in the research consisted of 101 data with 5 attributes consisting of age, Current Blood Sugar (GDS), 2 hours after eating/Post Pradial (PP), Fasting Blood Sugar (GPD) levels, and Low Density Lipoprotein (LDL) . To get classification results, there are several steps taken, namely data input, BE feature selection, 8-Fold Cross Validation, Naïve Bayes and results testing. From the classification results, testing was carried out using the accuracy, precision and recall calculation method. To find out the results of classification performance, four test scenarios were carried out, namely the first scenario, Naïve Bayes combined with BE and 8-Fold Cross Validation, accuracy of 77%, second scenario, Naïve Bayes combined with 8-Fold Cross Validation, accuracy of 78.1%, third scenario, Naïve Bayes combined with BE accuracy is 86% and the fourth scenario of Naïve Bayes classification accuracy is 90%, so the accuracy of Naïve Bayes classification with BE feature selection is better.Keywords: Diabetes melitus, Naïve bayes, Backward Elimination. 8-Flod Cross Validation.]]>
Copyrights © 2023
