Chronic Kidney Disease Prediction using Machine Learning Ensemble Algorithm

Abstract

Chronic Kidney Disease is one among the non-contagious illnesses that affect most of the individual in the world. The main factors of risk for the Chronic Kidney Disease are Diabetes, Heart Ailment, Hypertension. The Chronic Kidney Disease shows no symptoms in the early stages and most of the cases are diagnosed in the advanced stage. This leads to delayed treatment to the patient which may be fatal. Machine learning technique provides an efficient way in the prediction of Chronic Kidney Disease at the earliest stage. In this paper, four ensemble algorithms are used to diagnose the patient with Chronic Kidney Disease at the earlier stages. The machine learning models are evaluated based on seven performance metrics including Accuracy, Sensitivity, Specificity, F1-Score, and Mathew Correlation Coefficient. Based on the evaluation the AdaBoost and Random Forest performed the best in terms of accuracy, precision, Sensitivity compared to Gradient Boosting and Bagging. The AdaBoost and Random Forest also showed the Mathew Correlation Coefficient and Area Under the curve scores of 100%. The machine learning model proposed in this paper will provide an efficient way to prevent Chronic Kidney diseases by enabling the medical practitioners to diagnose the disease at an early stage.