Final Prediction Error of Autoregressive Model as a New Feature in the Analysis of Heart Rate Variability

The aim of this study is to offer a new heart rate variability (HRV) index that increases the accuracy in the discrimination of patients with congestive heart failure (CHF) from the control group. For this purpose, final prediction errors (FPE), which shows the quality of the conformity of autoregressive (AR) model, are calculated for model degrees from 1 to 100. Although the optimal AR model order and FPE values are widely used in the literature, they have not been used as possible HRV indices. In this study, we used FPE as an HRV feature for discriminating the patients with CHF from normal subjects and made a comparison with the other common HRV indices. As a result, we showed that FPE of AR model is a possible significant HRV feature.