Classification Model Based on Pathological Data for Kidney Diseases Prediction using Machine Learning Approach

Abstract

Chronic Kidney disease (CKD) is one of major threat all over the world with high morbidity and death rate. Patients often fail to diagnose the CKD problem as it lacks the symptoms at an early stage. Early identification of CKD can help us in reducing the death rate to a high extent as well as delays the further progression of disease. Machine learning models are built to predict the presence of CKD or not. Using the pathology data on the machine-learning model helps in detecting the CKD at an early stage. In this paper different classifiers such as KNN, Naïve Bayes, Random Forest, SVM and 3DCNN algorithms are compared for its predicting accuracy of CKD. 10 cross-validation techniques are sufficient for our model random forest, an ensemble approach which combines several decision tree models and their decision are combined to make the final prediction gives a maximum accuracy and precision in predicting the chronic kidney disease