EFSMLHA: Ensembled Feature Selected Machine Learning Hybrid Approaches for Heart Disease Prediction

The World Health Organization indicates that cardiovascular disease is the primary cause of death and disability globally. Preventive interventions, such as medicine and, in some cases, surgery, have saved millions of lives by reducing the prevalence of cardiovascular disease, made feasible by timely diagnosis. Machine learning and artificial intelligence have the capacity to significantly improve the precision of weather predictions, perhaps saving many lives. We have created a machine-learning model that integrates ten distinct machine-learning approaches to fulfill the need for early prediction. This collection includes techniques such as sequential feature selection (SFS), correlation-based feature selection (CFS), and recursive feature elimination (RFE). Hard voting, weighted voting, and soft voting are the three ensemble learning methodologies we have used. Our objective is to improve the prediction of heart illness by attaining elevated levels of specificity, accuracy, recall, and F1 score. We achieved our objective with an F1 score of 97.37 %, specificity of 94.44 %, recall of 97.37 %, accuracy of 96.43 %, and an area under the curve (AUC) of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{0. 9 8 3}$</tex>. Utilizing this model, medical personnel might more accurately predict the onset of cardiac disease, thereby facilitating earlier detection and intervention.