In the field of health, data mining can be used to predict a disease from patient medical record data, diabetes. There are several data mining models which one is classification. In the access field, there are many branches that are developing the decision tree (decision tree). One popular decision tree is C4.5. In this study, the data used were pima indian diabetes dataset taken from UCI machine learning repository. In this dataset all attributes are of continuous numeric type and for combined continuous data discretization is used. Accuracy is very important in the classification, ensemble method is a method used to improve the accuracy of classification algorithm by building some classifier of training data. From the research results, by applying discretization and bagging techniques to ensemble-based classification on C4.5 algorithm can increase the accuracy of 6.26%. With an initial accuracy of 68.61%, after applied discretization and bagging techniques to 74.87%..
Copyrights © 2017