Atrial fibrillation episodes detection based on classification of heart rate derived features

Atrial fibrillation (AF) is one of the most common cardiac arrhythmia and effects nearly 1-2 of every 100 persons of the population. This paper evaluates the effectiveness of Machine Learning (ML) approach to detect AF episodes. Features, determined exclusively on the basis of beat intervals, are classified with linear classifier. Performances of the proposed approach are evaluated by means of the MIT-BIH Atrial Fibrillation Database.