Myorcadial Infarction Classification with 15 Lead Electrocardiogram (ECG) Signal using Hybrid Convolutional Neural Network (CNN) and Bidirectional Long Short-Term Memory (BILSTM)

More than three branches of death due to cardiovascular disease. Myocardial infarction (MI) is one of the most dangerous forms of coronary heart disease, with the highest mortality rate among other types of coronary heart disease. MI is caused by reduced oxygen demand in heart muscle tissue when oxygen is pumped throughout the body. To detect MI, experts can use ECG signals. ECG is a tool that can be used to record the activity of biological signals formed due to the electrical activity of the heart. Various studies have been conducted using deep learning to help detect and predict MI. This research proposes a hybrid model of CNN-LSTM and CNN-BILSTM model techniques to predict MI. The MI dataset from the Physiobank (PTB) ECG database consisting of 52 normal subjects and 148 patients was used for training models. The proposed deep learning model uses automatic feature extraction. This study suggests a simple deep learning architecture with automatic feature extraction to learn from data. Evaluation of model performance resulted in an average accuracy of 98.80% for the CNN-LSTM model and 99.90% for the CNN-BILSTM model.