Application of C4.5 Classification Algorithm for Chronic Kidney Disease Diagnosis

Authors:

Abstract
The data from Ministry of Health Indonesia in the report infoDATIN 2017 mention that the Chronic kidney disease was the 27th leading cause of death globally then upward to 18th in 2010. In the report also states that chronic kidney disease prevalence increases with the rising in elderly population and also the increased cases of diabetes mellitus and hypertension. The difficulty of early detection is one reason for the study of the disease. The C4.5 algorithm is used to obtain a model for detecting chronic kidney disease. The classification model was built using 400 patient data as a dataset and 24 attributes in the form of symptoms suffered by patients. The C4.5 algorithm generates the decision tree so that it is simple to implement the classification model formed into the software with the ways a decision tree translated into classification rules. This algorithm was chosen because some other research in the same study case also shows the results that the C4.5 algorithm has an accuracy rate above 5O%. The results of the model testing produced by the C4.5 algorithm show the accuracy rate is 97.50% and a precision rate is 97.33% with a ROC curve (AUC) of 0.978. Thus, the software was built using the C4.5 classification model for diagnose Chronic kidney disease diagnosis has an excellent classification category.

This publication has 17 references indexed in Scilit: