Parameter Tuning Using Harris Hawks Optimization for Improved Chronic Kidney Disease Classification

At an early phase, chronic kidney disease (CKD) is usually not obvious. An appreciable reduction in kidney function is the primary sign of the disease. If CKD can be identified early, the rate at which CKD is advancing can be slowed down, and complications can be avoided. This study proposes using a swarm intelligence model for hyperparameter tuning. Hyperparameter tuning of SVM, RF, and GB classifiers was performed using the HHO Algorithm. The CKD dataset imbalance has been addressed by using SMOTE. Results of the study demonstrate that the HHO hyperparameter tuning methodology offered the best performance with respect to accuracy, F1-score, and ROC AUC. These results show that the HHO hyperparameter tuning methodology yields an increase in classification performance.