An Ensemble Learning Approach For Chronic Kidney Disease Classification

Abstract

Chronic kidney disease (CKD) is a potentially fatal condition that is difficult to identify early due to the absence of symptoms. The suggested study's goal is to create and verify a predictive model for chronic kidney disease prediction. It is continually rising into a worldwide health crisis. Unhealthy eating habits and a lack of water consumption are key reason to this disease. An individual can only survive for about 18 days without kidneys, necessitating kidney transplantation and dialysis. The present study proposes a method for forecasting CKD status using diagnostic medical data available on UCI repository that includes data pre-processing, a missing value management technique, aggregation of data, and extraction and classification. In this study, a variety of physiological parameters, as well as machine learning (ML) techniques such as train different machine learning models (SVM, KNN, Random Forest, Decision Tree, ADaBoost) with the normalized dataset In this study, we have purpose, a performance tuning nested approach is proposed that takes into account adjusting hyper - parameters as well as determining the appropriate weights to join ensembles (Ranking Weighted Ensemble)., with 98.75% accuracy, 100% Sensitivity, 96.55% Specificity and 99.03% f1score.. The findings suggest that it could be used to develop an automated system for detecting the severity of CKD.