Classifying cardiac biosignals using ordinal pattern statistics and symbolic dynamics

The performance of (bio-)signal classification strongly depends on the choice of suitable features (also called parameters or biomarkers). In this article we evaluate the discriminative power of ordinal pattern statistics and symbolic dynamics in comparison with established heart rate variability parameters applied to beat-to-beat intervals. As an illustrative example we distinguish patients suffering from congestive heart failure from a (healthy) control group using beat-to-beat time series. We assess the discriminative power of individual features as well as pairs of features. These comparisons show that ordinal patterns sampled with an additional time lag are promising features for efficient classification.