A novel APSGWO-XGBoost framework for heart disease detection: Enhancing diagnostic accuracy and efficiency in healthcare

Cardiovascular disease has been among the top leading causes of death in the world and timely diagnosis is essential to minimize death. Although machine learning (ML) has demonstrated potential in predicting heart diseases, the current models are usually affected by redundant features, poor hyperparameter optimization and premature convergence, which can limit the reliability of diagnostic results. In response to those issues, a new hybrid model, Adaptive Parallel Strategy Grey Wolf Optimizer with XGBoost (APSGWO-XGB), will be suggested in the current study. The framework has a beneficial selection process of features that helps in the reduction of 17 attributes to 9 discriminative clinical features that facilitates more interpretable results and helps clinician. Simultaneously, APSGWO hybrid optimizer combines both the global search power of Particle Swarm Optimization (PSO) and the local exploration power of the Grey Wolf Optimization (GWO) and adjusts exploration and exploitation in order to optimize XGBoost hyperparameters. It was run with the help of Python 3.10 and Scikit-learn and XGBoost libraries on the Kaggle Heart Disease Dataset (3000 records). The experimental outcomes indicate that the proposed framework is significantly better than the baseline models as it has an accuracy of 94.2% accuracy, 93.9% F1-score, 93.7% precision, 94.1% recall, and an AUC of 0.95. The proposed model improves accuracy (more than 86) and generalization with a difference of more than 8 % compared to conventional XGBoost (86 %) and single-optimizer models. These results point to the creativity and strength of combination of feature selection and adaptive hybrid optimization and a potent gradient-boosting classifier. This proposed APSGWO-XGB model offers a scalable, effective, and clinically explainable heart disease prediction tool, and has high potential of being used to support decision-making in real-time and early intervention in health care systems.