Nonlinear analysis of heart rate variability in the recognition of congestive heart failure

We report here studies of methods for recognizing early forms of congestive heart failure NYHA functional classes I–II based on application of nonlinear, statistical, and geometric parameters estimated using the Poincaré mapping to rhythm recordings. The studies showed that a linear classifier for identifying this pathology could be constructed and threshold decision rules were obtained. Experiments using real rhythm records selected from the PhysioNet web resource compared recognition quality obtained using a number of indicators.