Enhancing Cardiovascular Disease Prediction through Optimization Algorithms: A Comparative Study of Extra Trees Classifier and XGBoost Models

cardiovascular diseases (CVDs) continue to be a leading cause of death worldwide. To improve predictions and understanding of these conditions, we combined data from several well-known sources into one comprehensive dataset. We took an experimental approach, focusing on how different optimization algorithms affect the performance of two machine learning models: Extra Trees Classifier (ETC) and XGBoost (XGB). We used Particle Swarm Optimization (PSO), Monkey Search, Wolf Optimization, and Cuckoo Search to fine-tune the models. The data was scaled with MinMaxScaler and split into 70% training and 30% testing sets. We assessed the models based on accuracy, precision, recall, F1 score, and Cohen's kappa score, and also created ROC curves for a detailed performance analysis. Extra Trees Classifier (ETC) with Wolf Optimization yielded the best results, with an accuracy of 91%, precision of 94%, recall of 91%, F1 score of 92%, and a Cohen's kappa score of 82%. PSO, Monkey Search & Cuckoo Search, in contrast, resulted in lower performance metrics. XGBoost (XGB) with Wolf Optimization also led to impressive results with an accuracy of 89%, precision of 91%, recall of 89%, F1 score of 90%, and a Cohen's kappa score of 80%. The other optimization algorithms like PSO, Monkey Search & Cuckoo Search showed slightly less effective performance compared to Wolf Optimization.The findings highlight that Wolf Optimization significantly enhances the performance of both ETC and XGB models. This advancement is crucial for real-world applications, as accurate predictions of cardiovascular diseases can lead to early diagnosis and more effective management of the condition. By integrating these optimized models into clinical decision-support systems, healthcare providers can improve patient outcomes and resource allocation.