Enhancing kidney disease prediction with optimized forest and ECG signals data

Abstract

To improve the early detection of Chronic Kidney Disease (CKD) utilizing electrocardiogram (ECG) data, this study explores the use of the Optimized Forest (Opt-Forest) model. Exploiting the possible relationship between kidney function and ECG data, we investigate Opt-Forest's performance in comparison to popular machine learning (ML) models. We evaluate Opt-Forest and find that it outperforms other options in CKD prediction based on many measures such as classification accuracy (CA), false positive rate (FPR), and true positive rate (TPR). In comparison to previous models, Opt-Forest has superior sensitivity and specificity, with a TPR of 0.787 and a low FPR of 0.174. With an accuracy of 78.68 %, a KS of 0.641, and a low RMSE of 0.174, Opt-Forest also demonstrates robustness in CKD prediction. This study demonstrates the potential of Opt-Forest to improve patient outcomes and medical diagnostics, as well as the usefulness of ECG data in enhancing early CKD diagnosis. Prospective research avenues to advance precision medicine in nephrology involve investigating deep learning methodologies and incorporating patient-specific data.