Deep Learning-Based Heartbeat Classification of 12-Lead ECG Time Series Signal

Cardiovascular disease (CVD) is the most prevalent cause of mortality on a global scale. The possibility of preventing cardiac diseases is dependent on early-stage intervention. At this point, the electrocardiogram (ECG) emerges as a critical modality for detecting and predicting future cardiac abnormalities. Numerous machine-learning techniques have been used to improve result precision, computational efficiency, and model durability. Deep learning approaches have recently emerged as the preferred method for reaching even better levels of accuracy and robustness in a diverse range of applications. In this research, we developed four different neural networks. They are the Convolutional Neural Network (CNN), the Artificial Neural Network (ANN), the Long Short-Term Memory (LSTM) network, and a unique hybrid LSTM-CNN network. These networks were created to evaluate the classification performance between normal and abnormal heartbeats. Our hybrid model performed exceptionally well on the test data, with a loss of 0.12, binary accuracy of 0.97%, and area under curve (AUC) score of 0.99%. The results demonstrate its ability to successfully extract local and contextual factors within ECG signals. The combination of LSTM and CNN networks in a hybrid architecture allowed our model to exploit both local and contextual information within the ECG signals.