A DEEP LEARNING-BASED APPROACH FOR HEART DISEASE DETECTION AND EARLY WARNING

Abstract

As per the World Health Organization (WHO), 17.9 million individuals lose their lives annually owing to heart disease. With the help of image categorization, this research aims to utilize a deep learning approach. To identify heart problems. We utilized deep learning models including DenseNet201, ResNet-50, and CNN. The public is used to evaluate the proposed model. The heart disease dataset from the University of California, Irvine, consists of 1050 patients and 14 variables. The heart-disease dataset provided us with attainable features, and we considered this feature vector to be input for a dense net 201-used to determine whether an instance belongs to a healthy class or a class that is associated with heart illness. To determine the performance of the method that has been proposed, various performance indicators, including accuracy, precision, recall and F1score. Our proposed approach, densenet-201, achieved higher accuracy of 96%, with 0.91 precision, 0.93 recall and 0.96 f1 score. Finally we compared our approach with previous studies.