<![CDATA[Chronic kidney disease (CKD) is a progressive condition with an increasing incidence rate in Indonesia, including in Bojonegoro Regency. This disease results from a gradual decline in kidney function, leading to the accumulation of metabolic waste and toxins in the body. Early detection is crucial to prevent complications and enhance treatment effectiveness. However, current diagnostic methods rely heavily on laboratory tests and medical anamnesis, which may not provide sufficient accuracy. This study aims to develop a predictive model for kidney failure using Artificial Neural Network (ANN) and Random Forest (RF) algorithms. The ANN is particularly effective in recognizing complex nonlinear patterns, while RF demonstrates robustness in classification tasks. The results show that the ANN model achieved an accuracy of 81%, with blood urea nitrogen (BUN), blood pressure, creatinine, and diabetes identified as dominant predictors. Meanwhile, the RF model achieved an accuracy of 80%, with blood pressure emerging as the most influential variable. Based on comparative performance, ANN was selected as the optimal model for the kidney failure prediction. Nonetheless, integrating both algorithms into a hybrid framework could further enhance predictive coverage and support early detection and clinical decision-making in kidney failure management.]]>
Copyrights © 2025
