Comparative Study of Machine Learning Techniques for Prediction of Kidney Disease

Abstract

As kidney chronic disease is nowadays widely increasing which either caused by kidney disease or reduce the function of the kidney, it also affects the cardiac problems- scientifically which can lead to sudden heart attacks at the end-stage. Early diagnosis and adequate therapies can only help in stopping this disease, where dialysis and kidney transplantation is the only way to save the life of the patient. Detecting kidney disease through machine learning and through data mining techniques which can reveal the hidden problem of the kidney. Therefore, the current article is based on the comparative study using various Machine Learning techniques to detect kidney disease. This survey supports to find the accuracy of algorithms which are more useful to find the kidney disease in early stage. The comparative study of all the algorithms and by implementing the models on different platforms, and it is analyzed that which is the best algorithm to predict CKD (Chronic Kidney Disease). The machine learning techniques are compared like Probabilistic Neural Network (PNN), Multilayer Perceptron Algorithm (MLP), Logistic Regression (LOGR), Regression Tree (RPART), Support Vector Machine (SVM) and Radial Basis Function (RBF).