Early-stage diagnosis of chronic kidney disease using majority vote - Grey Wolf optimization (MV-GWO)

Chronic Kidney Disease (CKD) is a serious health issue that is growing at an alarming rate. A person having CKD can only be saved by providing kidney transplantation or dialysis but as it involves huge medical expenditure it remains unaffordable to low and middle-income countries. Therefore, early-stage diagnosis is critical to contain CKD at the initial stage, which can reduce the mortality rate and bring down the cost of treatment significantly. This study proposes a novel feature selection method namely Majority Vote-Grey Wolf Optimization (MV-GWO) with machine learning algorithms for the detection of CKD using benchmark CKD dataset. For validating the proposed method we have used different evaluation metrics. The proposed method achieved higher accuracy of 98.75%, sensitivity of 100%, F1-score of 99% and Mathew Correlation Coefficient of 0.9735 with only five optimal features. Furthermore, we have interpreted the machine learning model by analyzing the most critical features using SHapley Additive exPlanations (SHAP) feature importance plot, summary plot, dependence plot, and Partial dependence plots.