Automatic Detection of Atrial Fibrillation Using Spectrum Energy With Classified by Deep Learning CCN Types Adfal Afdala (a*), Nuryani Nuryani (b)
a) Sulthan Thaha Saifuddin Islamic State University of Jambi, Indonesia
*afdala[at]uinjambi.ac.id
b) Sebelas Maret University, Surakarta, Indonesia
Abstract
Atrial fibrillation is a heartbeat disorder that is hard to be predicted, yet has a quite dangerous effect, i.e the increase of stroke risk to its sufferers. The increasing number of these illness victims causes the need for an automatic detection system that is able to detect the symptoms of this heartbeat disorder, one of them is by using deep learning techniques. The feature used of this research is a spectrum of energy that is produced by the patient’s electrocardiogram signal. This research stage is started from data retrieval from the patients, signal processing, energy spectrum feature extraction, until the classification stage. This research produces a system that can detect atrial fibrillation from the patient’s electrocardiogram signal. The level of accuracy produced by this research reaches 92,50%. This result can be classified as good because it can be confirmed that the detection system produced in this research is able to detect atrial fibrillation well.
Keywords: Atrial Fibrillation, Deep Learning, Spectrum Energy
