Performance analysis of LSTM network in diagnosis of atrial fibrillation class of Arrhythmia

Abstract

The abnormality in the rhythm of the heartbeat is known as Arrhythmia. This abnormal rhythm can be fast or slow from the normal rhythm of the heartbeat. Certain abnormalities in the heart can be cured if the diagnosis is done at the earliest. The conventional method of recording and analyzing these rhythms are done using Electrocardiogram. This analysis is done manually which needs the expertise to identify the kind of heart disorder. An alternate solution for this problem is to implement the computer-assisted analysis. Machine learning methods are becoming more popular in various domains which makes a milestone every day through research and development. This paper discusses the results achieved with the implementation of a machine learning algorithm for the diagnosis of atrial fibrillation. The methodology used for the study is Long Short-Term Memory (LSTM) network to classify the signals used for the study into a normal and abnormal rhythm.