Hybrid ensemble learning model with bald eagle search feature optimization for heart disease diagnosis

Abstract

Heart disease is a wide phrase that covers a variety of disorders affecting the heart and blood arteries. It’s one of the leading causes of death globally. In some cases, heart disease develops silently over years, showing no symptoms until a major event like heart attack or stroke occurs. It results from a complex interplay of lifestyle, genetic, metabolic, and environmental factors. The growing prevalence and complexity of heart disease highlight the urgent need for intelligent, data driven early prediction systems. Healthcare, which has been transformed by machine learning, a branch of Artificial Intelligence, allows computers to learn and make precise predictions. Unlike traditional statistical techniques, Machine learning algorithms can model complex, non-linear relationships and capture subtle patterns in medical data, making them highly suitable for heart disease prediction. These models exposed to more data, so they improve over time, this makes them more accurate with continuous retaining. This work uses Bald Eagle Search optimization, a metaheuristic algorithm for feature selection. For classification, ensemble methods such as Random Forest, Voting, Bagging, Stacking, and various boosting techniques are used. The dataset for this work has 1025 instances and 14 features which is obtained from Kaggle. The result shows that Stacking has achieved highest accuracy of 95.61%.