An Intelligent System for Diagnosing Kidney Diseases Using Machine Learning

Early diagnosis plays a critical role in preventing severe complications and fatalities associated with kidney disease, especially in regions like Saudi Arabia, where healthcare access may be limited. Prompt identification of the disease can significantly mitigate its severity and enhance patient outcomes. This study introduces an intelligent system for early kidney disease diagnosis utilizing machine learning (ML) models. We conducted a comparative analysis of various models, such as Random Forest (RF), Logistic Regression (LR), Support Vector Classifier (SVC), Gradient Boosting (GB), K-Nearest Neighbor (KNN), and Decision Tree (DT), with the SVC model achieving the highest accuracy of 72.00%. RF and KNN closely followed with 68.00% accuracy each, while the DT model, despite its interpretability benefits, achieved an accuracy of 56.00%. The interpretability of the DT model could be particularly advantageous in healthcare diagnostics where transparency is crucial. RF exhibited the highest precision at 0.71, proving effective in reducing false positives. The proposed system presents a promising solution for improving early kidney disease diagnosis and could be seamlessly integrated into mobile platforms, offering practical real-world applications for healthcare professionals and patients, especially in underserved areas.