RETRACTED ARTICLE: A novel intelligent machine learning system for coronary heart disease diagnosis

Abstract

Coronary heart disease (CHD) is a significant medical disorder and one of the most prevalent forms of heart disease. Owing to the reality that a heart attack will happen without notice, an insightful screening system is inevitable. This paper investigates a new CHD detection approach built on an optimization machine learning technique, such as classifier ensembles. To boost the efficiency of our system, we used the Feature-Selector optimization model to select the best subset of CHD features. Second, to solve the problem of imbalanced CHD data, we used optimized SMOTE over-sampling, a highly efficient approach embedded with an optimization model. The class label estimation of three optimization learners, namely random forest, XGBoost API optimization, and SVM optimization model, is integrated in a stacked architecture. The identification model is validated using data from CHD patients. Finally, in terms of precision, F1, and ROC-Curve, our detection model outperformed existing ones focused on optimization models ensembles and individual classifiers. With random forest optimization, we achieved 90% accuracy, and with the XGBoost API optimization model, we achieved 89% accuracy. In contrast to previous reported research in the existing literature, this analysis indicates that our proposed model makes a substantial contribution.