This is a subset of the PTB-XL dataset. You will find in the file coorteeqsrfava.csv in the column 'ritmi' three diagnosis: NORMAL (label SR), ATRIAL FIBRILLATION (label AF) and all other VARIOUS Arrhytmias summed up into one class only (label VA). The Numpy file are the 12-leads ecg of the patients in 'coorteeqsrfava' file.

number of records 6528 ecg sampling 5000 (10 second recording at 500Hz) number of leads 12

The classes are composed on the rhythm diagnosis from PTB-XL:

SR 16782 sinus rhythm AFIB 1514 atrial fibrillation STACH 826 sinus tachycardia SARRH 772 sinus arrhythmia SBRAD 637 sinus bradycardia PACE 296 normal functioning artificial pacemaker SVARR 157 supraventricular arrhythmia BIGU 82 bigeminal pattern (unknown origin, SV or Ventricular) AFLT 73 atrial flutter SVTAC 27 supraventricular tachycardia PSVT 24 paroxysmal supraventricular tachycardia TRIGU 20 trigeminal pattern (unknown origin, SV or Ventricular)

Only 2000 records are taken from SR in order to build a more balanced datasets, they are sampled without repetition. AF is composed by AFIB and AFLT 1587 records. VA are all other types of arrhythmias summed up into one big class.