<![CDATA[Chronic Kidney Disease (CKD) is a serious global health issue, ranking as the 12th leading cause of death in 2019, with a 31.7% increase since 2010. Many CKD patients require hemodialysis, which poses risks of complications such as hypertension, hypotension, and gastrointestinal disorders, increasing mortality. This study predicts hemodialysis complications using XGBoost optimized with the Artificial Bee Colony (ABC) algorithm. The dataset includes numerical and categorical variables such as blood pressure, hemoglobin levels, gender, and complication history. To improve class distribution, the Synthetic Minority Over-sampling Technique is applied. Five test scenarios with different ABC parameter configurations were conducted to optimize XGBoost hyperparameters. Results indicate that balancing the dataset with SMOTE enhances model accuracy. Among the tested scenarios, Test 3, with ABC parameters n_bees set to 30, max_iter set to 30, and limit set to 10, achieved the highest accuracy, increasing from 89% (unbalanced) to 94% (balanced). Although training time increased, the improved performance highlights the potential of the XGBoost-ABC framework for early complication detection. This approach can enhance patient care, reduce mortality risks, and support clinical decision-making for hemodialysis patients.]]>
Copyrights © 2025
