Hybridizing Wolf Search Algorithm With Xgboost Model For Accurate Identification Of Cardiac Disorders

Abstract

Abstract. Rapid and accurate diagnosis is necessary to treat heart disorders, a global health issue. Medical diagnosis tasks have shown promising results with machine learning, notably ensemble algorithms like XGBoost. Still, you must adjust these models’ hyperparameters to maximize their performance. This research combines the Wolf Search Algorithm (WSA) with XGBoost to improve heart disease identification. We apply WSA to optimize XGBoost classifier hyperparameters like learning rate, tree depth, and regularization. A big collection of diagnostic and clinical data from patients with various cardiac conditions was used for our tests. Preprocessing addressed missing val- ues and ensured uniform scaling. Our hybrid methodology was tested using rigorous cross-validation methods to determine AUC-ROC, sensitivity, specificity, and accuracy. A combo of WSA and XG- Boost enhances heart issue diagnosis accuracy compared to conventional parameter tuning methods. The poposed model gained 0.973 accuracy level, 0.97 precision value, 0.89 recall value with 0.93 f1-score. Several performance indicators show the upgraded XGBoost model can discriminate car- diac conditions. Additional insights on model interpretability and feature importance for diagnostic decision-making are offered. We found that XGBoost and swarm intelligence algorithms like WSA can increase heart disease diagnosis reliability and accuracy. Implementing the provided methods in clinical settings may improve healthcare outcomes and patient management.