Comparative Analysis of Predictive Machine Learning Algorithms for Chronic Kidney Disease

Abstract

Chronic Kidney Disease (CKD) is a serious lifelong ailment caused kidney malfunction. To diagnose CKD, number of EHRs generated and manually process them is difficult. So, to process these EHRs and to diagnose CKD Machine learning approaches can assist. The main motive of this study is to use machine learning algorithms to diagnose CKD in order to assist people to avoid it. The CKD Dataset acquired from the UCI machine learning repository is analyzed using machine learning techniques such as Support Vector Machine (SVM), Naive Bayes (NB), Decision Tree (DT), Artificial Neural Network (ANN), Random Forest (RF), K-Nearest Neighbor (K-NN), and Logistic Regression (LR). WEKA v3.8.3 was used to conduct the experiment. The mean of the values was used to replace missing values. With 99 % accuracy, RF outperforms other approaches. Against the CKD dataset, the ROC curves of LR and ANN yielded 0.999 and 0.998 results, respectively. This research can assist healthcare institutions and researchers in predicting CKD at an early stage.