Effective Feature Engineering Technique for Heart Disease Prediction With Machine Learning

Abstract

Heart failure is a chronic disease affecting millions worldwide. An efficient machine learning-based technique is needed to predict heart failure health status early and take necessary actions to overcome this worldwide issue. While medication is the primary treatment, exercise is increasingly recognized as an effective adjunct therapy in managing heart failure. In this study, we developed an approach to enhance heart failure detection based on patient health parameter data involving machine learning. Our study helps improve heart failure detection at its early stages to save patients’ lives. We employed nine machine learning-based algorithms for comparison and proposed a novel Principal Component Heart Failure (PCHF) feature engineering technique to select the most prominent features to enhance performance. We optimized the proposed PCHF mechanism by creating a new feature set as an innovation to achieve the highest accuracy scores. The newly created dataset is based on the eight best-fit features. We conducted extensive experiments to assess the efficiency of several algorithms. The proposed decision tree method outperformed the applied machine learning models and other state-of-the-art studies, achieving a high accuracy score of 100%, which is admirable. All applied methods were successfully validated using the cross-validation technique. Our proposed research study has significant scientific contributions to the medical community.

Keywords

• Machine learning
• heart failure
• cross validations
• feature engineering