Predicting chronic kidney failure disease using data mining techniques

Abstract

Kidney failure disease is being observed as a serious challenge to the medical field with its impact on a massive population of the world. Devoid of symptoms, kidney diseases are often identified too late when dialysis is needed urgently. Advanced data mining technologies can help provide alternatives to handle this situation by discovering hidden patterns and relationships in medical data. The objective of this research work is to predict kidney disease by using multiple machine learning algorithms that are Support Vector Machine (SVM), Multilayer Perceptron (MLP), Decision Tree (C4.5), Bayesian Network (BN) and K-Nearest Neighbour (K-NN). The aim of this work is to compare those algorithms and define the most efficient one(s) on the basis of multiple criteria. The database used is “Chronic Kidney Disease” implemented on the WEKA platform. From the experimental results, it is observed that MLP and C4.5 have the best rates. However, when compared with Receiver Operating Characteristic (ROC) curve, C4.5 appears to be the most efficient.