Employing Convolutional Neural Network Model in Predicting Heart Diseases

Heart disease is a global health challenge and one of the leading causes of death worldwide. Early diagnosis of heart disease allows for appropriate therapeutic intervention and necessary measures to prevent disease progression and prevent serious complications. The significance of ECG lies in its ability to detect heart disease and accurately diagnose various conditions. Many research studies have been conducted in the field of artificial intelligence techniques to analyze cardiac signals. This study aims to create a one-dimensional convolutional neural network (CNN) to process these ECG signals utilizing MIT-BHI database that contains 87554 samples of ECG signals was used. The database contains different cardiac signal classes, including normal signals, regular atrial contraction, premature ventricular contraction, combined beat, which results from the superposition of several waves in the presence of a cardiac pacemaker, and a class labeled as "unknown signals". Data were divided after processing into 80% training data, 15% test data, and 5% validation data. We processed the data and trained the CNN neural network model accordingly, and the results showed that the diagnostic accuracy reached 99.52% for training and 99.45% for testing. Keywords: Deep Learning, Covolutional Neural Network, Heart Disease Prediction, Healthcare.