Comparative Analysis of Datamining Algorithms for Heartbeat Level Prediction

Though irregular heart rates can suggest various cardiac and other health ailments, accurately forecasting specific heartbeat patterns holds a crucial role in medical diagnoses and devising treatment strategies. This study introduces a deep learning-grounded method aimed at discerning between normal and pathological readings in electrocardiograms (ECGs). The approach employs stacked denoising autoencoders (DAEs) and deep neural network (DNN) classifiers, meticulously trained on an extensive dataset of ECG recordings. The evaluation encompasses diverse metrics, including accuracy, precision, recall, and the F1-score. A comparative analysis is undertaken with advanced approaches such as support vector machines, back-propagation neural networks, and general regression neural networks. The experimental findings compellingly showcase the superiority of our proposed approach across a range of criteria, revealing outstanding performance. For instance, we attain an impressive score of 95.2%, an overall reliability rate of 95.2%, accuracy elevated to95.4%, recall peaking at 95.0%, and a noteworthy ROC curve performance of 98.9%. In conclusion, the deep learning-based approach introduced in this study holds substantial potential for accurately and effectively predicting heart conditions. Its potential applications in healthcare settings encompass early detection and adept management of cardiac issues and other health-related considerations.