Heart Failure Detection Using Deep Neural Network

An estimated million people die from heart disease each year, which makes it the most common cause of death in the world. Heart failure can be predicted early, which can increase mortality rates. In this context, numerous classification techniques are described in the literature. As a result, it has turned into a difficult issue in the healthcare domain. In this paper, a deep learning-based classifier is proposed for heart failure detection. Additionally, pre-processing steps like normalization and feature encoding strategies are introduced in a novel way to enhance the performance of the suggested classifier. To eliminate the struggle of hyper-parameters selection, grid-search is used. To show the robustness of the proposed classifier, a standard heart disease dataset is utilized, and results are obtained under various training-validation-testing ratios. The findings show that the performance of the proposed classifier is better than the methods currently in use for classifying heart disease. The suggested classifier might be employed as a tool to help experts make early predictions about heart disease.