Diagnosis of Coronary Heart Disease by Combination of Random Forest and Grey Wolf Optimization Algorithms

In recent years, cardiovascular disease has become one of the leading causes of death worldwide. Accurate identification of risk factors is crucial for early diagnosis and effective management of this disease. In this paper, we propose a hybrid approach that combines the Grey Wolf Optimization (GWO) algorithm for feature selection with the Random Forest (RF) classifier (GWO-RF) to diagnose coronary heart disease using the Cleveland Heart Disease dataset. The GWO algorithm is employed to select the optimal subset of features, removing redundant attributes and enhancing model performance, while RF performs the classification task. All models were assessed using a 5-fold cross-validation strategy. In addition to the proposed GWO-RF model, GWO-based feature selection was also applied to Logistic Regression (GWO-LR) and K-Nearest Neighbors (GWO-KNN) for comparative analysis. The experimental results demonstrate that GWO-RF achieves an accuracy of 94.38%, an F1-score of 93.66%, and an AUC of 97.91%, outperforming the compared models. These findings confirm that GWO-based feature selection effectively improves classifier performance and enhances predictive reliability in coronary heart disease diagnosis.