RR interval time series modeling: the PhysioNet/Computers in Cardiology challenge 2002

Cardiac inter-beat (RR) interval time series contain uc-tuations at time scales ranging from a few seconds to many hours. Realistic models of these series are potentially use-ful to researchers, not only as sources of surrogate data with known properties for evaluating novel analytic meth-ods, but also as sources of insight into the diverse mech-anisms underlying heart rate variability. PhysioNet and Computers in Cardiology have sponsored an open on-line competition aimed at stimulating the creation and exchange of high-quality models of RR interval variability, the third in an annual series of challenges for the research community. Participants rst created software that was used to generate 24-hour synthetic time series, then attempted to identify the synthetic series within an unlabeled data set that included roughly equal amounts of real and synthetic data. All of the software models and the data used in the challenge are available a