<![CDATA[Diabetes is a deadly and chronic disease. It characterized by an increase in blood sugar. Many complications occur if diabetes does not treat and identified. The common identification process by visits to diagnostic centers and consulting physician. It makes bored patients. Machine learning approach can solve the problem of diabetic identification. However, the unbalanced range of diabetes variable values ​​affects the quality of machine learning results. This study predicts the likelihood of diabetes in diabetic patients from 768 Indian women, using three machine learning classification algorithms and Z-Score normalization method. The machine learning algorithms used are Decision Tree, Support Vector Machine (SVM) and Naive Bayes. Experiments were run on the Pima Indians Diabetes Database (PIDD). Dataset retrieved from the UCI Machine Learning Repository. The performance of the three algorithms was evaluated using accuracy, precision, F1, and recall based on confusion matrix. SVM algorithm is an algorithm that has the highest performance that both algorithm the Naive Bayes and Decision Tre algorithms, the accuracy and F1 is 80.73% and 76%. The Z-Score method has positively contribution to increasing the accuracy of the classification model. Furthermore, this study also managed to get a higher accuracy than previous studies.]]>
Copyrights © 2022
