Abstract
Atrial Fibrillation (AF) is a prevalent form of arrhythmia that
primarily affects the atria (upper chambers of the heart), causing
disorganized electrical signals and resulting in an irregular heartbeat.
Electrocardiography (ECG) is the most commonly used technology for
detecting AF, as ECG signals provide crucial information on cardiac
irregularities, making it an essential tool for AF diagnosis. However,
manually analyzing ECG signals is both mentally demanding and
time-consuming. In this study, we propose Hybrid CNN-LSTM model to
determine whether a given ECG signal represents a normal sinus rhythm or
indicates Atrial Fibrillation. We utilize the Physionet Database for our
dataset. The proposed hybrid model aims to assist doctors by
significantly reducing the time and mental effort required for
diagnosis, facilitating early detection of the disease and lowering the
risk of thromboembolic events, including ischemic strokes. Our Hybrid
CNN-LSTM model is designed to autonomously learn discriminative features
from raw ECG signals, eliminating the need for manual feature
extraction. We compared the performance of the Hybrid CNN-LSTM model
with a standard CNN model. The Hybrid CNN-LSTM model achieved an
accuracy of 95.2%, outperforming the CNN model, which achieved an
accuracy of 81.4%. Therefore, our proposed model offers a more accurate
method for