The PAF database are given here.
Please randomly split it into training set and test set. The sampling rate for the PAF database is 128 samples/sec.
The p* learningset is people with confirmedPAF. The variablesof type p* that endin odd numbersare ecg recordings thatcontain NO PAF. Thefollowingeven p*record is theecg recording immediately before anepisode ofPAF.The variables thatendin "c", area 5 minutecontinuationofthe 30minuterecording. Thus,in an odd numberedp-variable, p*c shows NOPAF and aneven numbered variable shows PAF.
Example:p01 and p01c arerecords that containnoPAF. p01 is 30 minutes and p01c is 5 minutes. p02 is the 30 minute recording from the same subject(p01) and is thetime immediately beforean episode ofPAF. p02c is the 5 minute continuation from p02 and contains PAF.
The n* learning set isnumbered and organized inthe same way is the p* test set, except that all of the n* records are of people confirmed with NO PAF.
Possiblefeaturescould beusedinclude spectrograms of the heart rate signals, signal-averaged P-wave morphology, R-R interval dynamics, and atrial ectopy. For moreexplanation, you can visit http://www.physionet.org/physiobank/database/afpdb/
R-R Interval Features: Download Here.
The unzipped .ihr files could be read using emacs or other text editor. The first two columns in these files correspond to seconds (marked where the qrs is detected) and the nstantaneous heart rate (1/ibi).
The records with names beginning with p come from subjects who have paroxysmal atrial fibrillation (PAF). The second (even-numbered) record in each pair of 30-minute records contains the ECG immediately preceding an episode of PAF.
The records with names
beginning with n come from subjects who do not have documented atrial fibrillation,
either during the period from which the records were excerpted or at any
other time. The subjects include healthy controls, patients referred for
long-term ambulatory ECG monitoring, and patients in intensive care units.