A Bayesian classification of heart rate variability data

We propose a simple Bayesian method for the classification of time series signals originating from mutually exclusive sources. In particular, the method is used to address the question of whether a 24-h recording of human heart rate data is produced by a normally functioning heart or by one exhibiting symptoms of congestive heart failure. Our method correctly classifies 18 of 18 normal heart data sets, and 38 of 44 congestive failure data sets.