Chronic Kidney Disease Prediction using Ensemble Machine Learning

Abstract

Abstract: Chronic kidney disease is considered one of the major diseases now-a-days. Most of the people are affected for their irregular lifestyle. Early-stage prediction can reduce it and can suggest a healthy lifestyle. In this study, we predict kidney disease from secondary data using some machine learning algorithms. Support Vector Machine (SVM), Random Forest (RF), Logistic Regression (LR), K Nearest Neighbors (KNN), Stotostical Gradient Deasent (SGD) are used for analysis. We also propose an ensemble machine learning algorithm by stacking RF, SVC, and LR and named RFSVCLR. This algorithm shows better result than others classifiers. Precision, Recall, F1 Score, Accuracy, Cohen Kappa, and ROC is used to evaluate the performance of the algorithms. RFSVCLR shows 99% accuracy with 99% precision, 99% recall, 99% f1 score and 98% Cohen kappa score that is superior to other classifiers.