WITHDRAWN: Harnessing deep learning for accurate heart disease prediction

Abstract Heart disease is leading global health challenge, contributing significantly to mortality rates. Accurate detection of heart diseaseis vital for timely intervention. This research paper explores the development and evaluation of an effective deep learning (DL)model for the accurate detection of heart disease. The proposed model leverages the power of deep learning, specificallydeep neural networks (DNNs), to analyze a comprehensive Cleveland dataset containing 303 instances with 14 attributes andthe heart disease dataset comprising 1025 instances with 14 attributes, ensuring the model’s robustness and applicabilityacross different demographics. We present a feature engineering approach to preprocess the raw data effectively, ensuring thatrelevant information is extracted and utilized by the deep learning model. The performance of the presented model is evaluatedin terms of accuracy, precision, recall, F1-score, Cohen’s kappa, receiver operating characteristics (ROC), area under thecurve (AUC), and specificity. For the Cleveland dataset, the model got 99.06% accuracy, 100% precision, 98.03% recall, 99%F1-score, AUC of 0.99, and 100% specificity. Similar to this, for the heart disease dataset, the model got 100% accuracy, 100%preci-sion, 100% recall, 100% F1-score, 100% AUC, and 100% specificity, which is superior to existing research that usedfeature selection (FS) techniques in their research. The experimental findings of the work indicate that the presented methodoutperformed the existing methods with remarkable performance.